INDEX

Explanations

proper nouns mentioning different sources or organizations

oai_token-act-pair · gpt-3.5-turbo

references to sources or authorities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

abiding

-0.72

 mattered

-0.68

rontal

-0.64

ront

-0.63

obyl

-0.63

btn

-0.62

atible

-0.62

NetMessage

-0.61

ezvous

-0.60

 confronts

-0.60

POSITIVE LOGITS

 datas

0.85

 estimates

0.80

reports

0.76

 lore

0.74

 calculations

0.71

 estim

0.71

 folklore

0.69

 sources

0.69

 specifications

0.69

 quoting

0.69

Activations Density 0.287%

proper nouns mentioning different sources or organizations

references to sources or authorities

No Comments

No Known Activations