Run #758

done rubric v1 session 79123284 gemini-3.7-flash 2026-09-21 02:21

What the student sent the question as it arrived, then each reading of it

OCR of the image GOOGLE_VISION

9:33 PM ποιο παπ Meta Al❤ 4G 68 B.Sc Chemistry Hons - 3 saal / 6 Semester NEP ke hisaab se 4 saal bhi ho sakta hai 1. Core Chemistry Subjects - Har semester me honge Inorganic Chemistry - Periodic table, Coordination compounds, Chemical bonding Organic Chemistry - Reaction mechanism, Biomolecules, Spectroscopy Physical Chemistry - Thermodynamics, Chemical kinetics, Quantum chemistry, Electrochemistry 2. Allied/Subsidiary Subjects Message

The phone's own reading, which differs:

9:33 PM ্র Meta AI ।।७।ig। B.Sc Chemistry Hons - 3 saal /6 Semester NEP ke hisaab se 4 saal bhi ho sakta hai 1. Core Chemistry Subjects - Har semester me honge Inorganic Chemistry - Periodic table, Coordination compounds, Chemical bonding Organic Chemistry - Reaction mechanism, Biomolecules, Spectroscopy Physical Chemistry - Thermodynamics, Chemical kinetics, Quantum chemistry, Electrochemistry 2. Allied/Subsidiary Subjects Message ६

The audit's reading of the question, which is what the Product AI judged against:

B.Sc Chemistry Hons - 3 saal / 6 Semester NEP ke hisaab se 4 saal bhi ho sakta hai 1. Core Chemistry Subjects - Har semester me honge Inorganic Chemistry - Periodic table, Coordination compounds, Chemical bonding Organic Chemistry - Reaction mechanism, Biomolecules, Spectroscopy Physical Chemistry - Thermodynamics, Chemical kinetics, Quantum chemistry, Electrochemistry 2. Allied/Subsidiary Subjects

Audit's summary: The student asked for information about the B.Sc Chemistry Hons syllabus, specifically for the 3rd year/6th semester. The tutor provided the syllabus details, including core and allied subjects, and explained that the NEP (National Education Policy) might extend the program to 4 years. The tutor also confirmed that the syllabus is primarily in Hindi with some English terms.

Transcript 505 characters · 150 stripped

Read it no speaker labels — tutor and student run together

Hello Anurag, welcome to Filo. Anurag, am I audible? Yes. Anurag, बताओ क्या जाओ चाहूँ? Yes, वे. English ज़िखा तो हिंदी करके बताए तो. ठीक है, just give a second. आपका नाम? Okay, दोस्तों, आवाज आ रही है ना. Hello. आपको बीस्ती ओनर्स के वाइंग बताया जाये आवाज आ रही है? नो Hello. Your voice is cutting. One minute, wait. Your voice is coming. Yes, your name? My name is Sakshi. Sakshi, is it? Okay. Look, what is written here? PhD, Honours, 3 years and 6 semesters. बस अनुराग इतना ही लिखा है यहाँ पर. Thank you.

What the student sent the question as it arrived, then each reading of it

OCR of the image GOOGLE_VISION

The phone's own reading, which differs:

The audit's reading of the question, which is what the Product AI judged against:

Transcript 505 characters · 150 stripped

Read it no speaker labels — tutor and student run together

Review team

no_issue

Product AI

nothing found

teaching Yes · mistakes No · satisfactory Yes

Each verdict takes your own. Yours is stored against the stack's, so “where is it wrong most often” is a question with an answer.

Work is the right way up Canvas is upright

The screenshot content is oriented upright at 0 degrees across all frames (Frames 1-8).

seen · 100% confident
Your view
Writing is legible at video size Typed text on image is clear and legible

The uploaded screenshot contains clear typed digital text in black on a white background (Frames 1-8).

seen · 100% confident
Your view
The work fills the frame Not applicable to a board session

seen
Nothing distracting in shot Not applicable to a board session

seen
The student's problem is established The student's material is visible from the beginning

The screenshot regarding B.Sc Chemistry Honours syllabus details is visible from Frame 1.

seen · 100% confident
Your view
Work proceeds in followable steps Tutor systematically annotates the provided text

The tutor highlights and checks off each point on the screenshot progressively (Frames 4 to 8).

seen · 90% confident
Your view
It reaches an answer The text explanation was completed

The tutor went through all the listed core chemistry subjects and concluded by reading all visible items (Frames 6-8 and transcript).

seen · 90% confident
Your view
Something was actually written board went from 16.05% to a peak of 31.25% covered (+15.20 points), ending at 17.22%

measured
Your view
The concept is named before the working No academic concept named

The session involves general syllabus reading/translation without naming an academic concept.

heard · 90% confident
Your view
Explained rather than jumped to the answer Tutor goes directly into reading the text

After asking the student's preference, the tutor immediately starts reading the text provided in the screenshot.

heard · 90% confident
Your view
Solved step by step, not in one leap Tutor goes through the points sequentially

The points are highlighted one by one on the screen in sequential order (Frames 4 to 8).

heard · 90% confident
Your view
The session opens properly Tutor greets and confirms student requirement

Transcript shows: 'Hello Anurag, welcome to Filo. Anurag, am I audible?... बताओ क्या जाओ चाहूँ?'

heard · 100% confident
Your view
No long dead air 0 gaps over 6.0s, 0s total (0% of the session), longest 0s

Silence is below -35.0dB — a crude measure on a phone microphone, so treat a low count as weak evidence rather than proof the tutor was talking.

measured
Your view

Observations judged, but not counted as the lab objecting

The camera holds still Not applicable to a board session

Not counted: Measured over 183 known-bad and 29 known-good camera sessions (2026-09-21): raised on 87% of the bad and 86% of the good, and the motion distributions are the same shape — median frame difference 25 on the bad set, 27 on the good. No threshold on this number separates them, so it is reported and not counted.

measured
Exposure is usable Not applicable to a board session

Not counted: Same two sets: it raised 6 of 183 known-bad and 2 of 29 known-good sessions. Both too few to mean anything, and the brightness distributions overlap almost exactly. The earlier reading that it 'fires more on clean sessions' was two sessions out of twenty-nine — noise, not an inverted rule.

measured
Understanding of the concept is checked No conceptual questions asked

The tutor only checks audio connectivity and reads the text without asking any concept-checking questions.

Not counted: Raised on 87% of known-bad and 80% of known-good sessions. The criterion is honest about why in its own `why`: the transcript has no speaker labels, so a real check and a rhetorical 'ok?' are often indistinguishable, and the bar ends up one almost no session clears. Still judged and shown, because reading it beside a recording is how it gets better.

heard
Your view

What the model returned verbatim

{
  "findings": [
    {"key": "orientation", "verdict": "ok", "headline": "Canvas is upright", "evidence": "The screenshot content is oriented upright at 0 degrees across all frames (Frames 1-8).", "confidence": 1.0},
    {"key": "legibility", "verdict": "ok", "headline": "Typed text on image is clear and legible", "evidence": "The uploaded screenshot contains clear typed digital text in black on a white background (Frames 1-8).", "confidence": 1.0},
    {"key": "question_captured", "verdict": "ok", "headline": "The student's material is visible from the beginning", "evidence": "The screenshot regarding B.Sc Chemistry Honours syllabus details is visible from Frame 1.", "confidence": 1.0},
    {"key": "structure", "verdict": "ok", "headline": "Tutor systematically annotates the provided text", "evidence": "The tutor highlights and checks off each point on the screenshot progressively (Frames 4 to 8).", "confidence": 0.9},
    {"key": "completion", "verdict": "ok", "headline": "The text explanation was completed", "evidence": "The tutor went through all the listed core chemistry subjects and concluded by reading all visible items (Frames 6-8 and transcript).", "confidence": 0.9},
    {"key": "concept_named", "verdict": "bad", "headline": "No academic concept named", "evidence": "The session involves general syllabus reading/translation without naming an academic concept.", "confidence": 0.9},
    {"key": "explained_before_solving", "verdict": "weak", "headline": "Tutor goes directly into reading the text", "evidence": "After asking the student's preference, the tutor immediately starts reading the text provided in the screenshot.", "confidence": 0.9},
    {"key": "concept_checked", "verdict": "bad", "headline": "No conceptual questions asked", "evidence": "The tutor only checks audio connectivity and reads the text without asking any concept-checking questions.", "confidence": 1.0},
    {"key": "step_by_step", "verdict": "ok", "headline": "Tutor goes through the points sequentially", "evidence": "The points are highlighted one by one on the screen in sequential order (Frames 4 to 8).", "confidence": 0.9},
    {"key": "opening", "verdict": "ok", "headline": "Tutor greets and confirms student requirement", "evidence": "Transcript shows: 'Hello Anurag, welcome to Filo. Anurag, am I audible?... बताओ क्या जाओ चाहूँ?'", "confidence": 1.0}
  ]
}

Duration	158s (2.6 min)
Video	1280×800 @ 10.0fps
Audio track	yes
Shot changes
Dead air	0 gaps ≥6.0s · 0s total · longest 0.0s
Luma	avg 89.1 · min · max
Medium	board 73.7% flat background
Content written	16.05% → 17.22% (+15.2)
Transcript	505 characters of speech
Frames	8

Run #758

The recording

What the student sent the question as it arrived, then each reading of it

Transcript 505 characters · 150 stripped

What the student sent the question as it arrived, then each reading of it

Transcript 505 characters · 150 stripped

Review team

Product AI

Measured

What the model returned verbatim

Frames in order, one every ~20s