INDEX

Explanations

sentence structure of interview transcripts

oai_token-act-pair · gpt-3.5-turbo

instances of numerical data or significant figures within texts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 estranged

-0.88

 repatri

-0.77

nab

-0.75

 tempted

-0.73

 illicit

-0.71

 await

-0.71

 unexplained

-0.69

 unin

-0.69

 formally

-0.69

 suspected

-0.69

POSITIVE LOGITS

Interview

1.07

Narr

1.05

Anyway

1.01

Newsletter

1.00

laughs

0.96

 GOODMAN

0.95

Looking

0.94

Then

0.94

Talking

0.92

Certainly

0.92

Activations Density 0.285%

sentence structure of interview transcripts

instances of numerical data or significant figures within texts

No Comments

No Known Activations