INDEX

Explanations

phrases related to negative events or controversies

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

terms related to government accountability and public policy issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tis

-0.62

tin

-0.61

Rating

-0.60

 commits

-0.59

Tok

-0.59

 participates

-0.59

ieth

-0.58

 delivers

-0.57

Christmas

-0.56

 realizes

-0.56

POSITIVE LOGITS

 abound

1.36

 vary

1.20

 arise

0.93

 continue

0.92

 differ

0.91

 emerge

0.90

 remain

0.90

 emerged

0.90

 arose

0.89

 have

0.89

Activations Density 0.201%

phrases related to negative events or controversies

terms related to government accountability and public policy issues

No Comments

No Known Activations