INDEX

Explanations

statements related to belief or trust

oai_token-act-pair · gpt-3.5-turbo

phrases that express belief or trust

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

imilar

-0.77

aida

-0.71

idian

-0.70

 sidel

-0.66

entin

-0.66

undy

-0.66

inav

-0.65

imer

-0.65

ertation

-0.65

ija

-0.63

POSITIVE LOGITS

 when

0.72

 WHEN

0.65

 Yourself

0.61

 hype

0.60

 yourselves

0.57

 unless

0.57

 expr

0.56

 whenever

0.56

 eminent

0.56

}}

0.55

Activations Density 0.069%

statements related to belief or trust

phrases that express belief or trust

No Comments

No Known Activations