INDEX

Explanations

phrases related to agreements being officially endorsed or approved, often involving countries

oai_token-act-pair · gpt-3.5-turbo

terms related to official approvals or legal agreements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ster

-0.82

fen

-0.80

hunt

-0.74

visor

-0.74

eter

-0.73

ter

-0.73

nut

-0.72

userc

-0.72

uber

-0.71

crow

-0.69

POSITIVE LOGITS

 ratified

0.93

 Seym

0.89

 GOODMAN

0.82

externalActionCode

0.80

é¾įå¥ĳå£«

0.73

 Cosponsors

0.73

ALLY

0.72

wana

0.70

 Votes

0.70

 Sabha

0.69

Activations Density 0.019%

phrases related to agreements being officially endorsed or approved, often involving countries

terms related to official approvals or legal agreements

No Comments

No Known Activations