INDEX

Explanations

media-related terms such as show, movie, news, and talk

oai_token-act-pair · gpt-3.5-turbo

references to media and entertainment, particularly related to talk shows and news

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

TPPStreamerBot

-0.77

 Helpful

-0.74

yss

-0.72

 individually

-0.66

 describ

-0.66

 Continuous

-0.65

Specific

-0.65

 downstream

-0.63

 Effect

-0.62

 Roof

-0.62

POSITIVE LOGITS

 outfit

0.83

warts

0.77

aunts

0.77

lance

0.75

coon

0.75

 enthusiast

0.74

 theorist

0.73

abilia

0.72

 guru

0.72

duo

0.71

Activations Density 0.518%

media-related terms such as show, movie, news, and talk

references to media and entertainment, particularly related to talk shows and news

No Comments

No Known Activations