INDEX

Explanations

references to research studies or reports

oai_token-act-pair · gpt-3.5-turbo

references to studies and reports in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

å§«

-0.77

à¨

-0.68

itely

-0.59

 rebuilt

-0.59

uxe

-0.58

ply

-0.53

udeau

-0.52

 toget

-0.52

isable

-0.52

 intervals

-0.51

POSITIVE LOGITS

 regarding

0.92

 titled

0.88

 cited

0.82

 entitled

0.81

cerning

0.80

 pertaining

0.79

 concerning

0.79

 announcing

0.78

 confirming

0.76

 condemning

0.76

Activations Density 0.263%

references to research studies or reports

references to studies and reports in the text

No Comments

No Known Activations