INDEX

Explanations

mentions of talk-related activities or professions

oai_token-act-pair · gpt-3.5-turbo

references to talk shows and discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

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

 Unsure

-0.62

 Reserve

-0.60

cil

-0.60

 Cust

-0.60

aples

-0.59

————

-0.58

 Swan

-0.57

 Soldier

-0.56

wcs

-0.56

 handc

-0.56

POSITIVE LOGITS

ative

1.20

radio

1.03

back

0.96

athon

0.92

Talk

0.91

about

0.91

show

0.89

ership

0.87

backs

0.86

talk

0.83

Activations Density 0.038%

mentions of talk-related activities or professions

references to talk shows and discussions

No Comments

No Known Activations