INDEX

Explanations

information related to political figures and discussions about political strategies and alliances

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

mentions of government officials and positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

fif

-0.75

uts

-0.70

ives

-0.69

ahon

-0.68

otton

-0.66

everal

-0.66

hedon

-0.65

andals

-0.63

flix

-0.63

uga

-0.63

POSITIVE LOGITS

 Trek

0.62

 Derby

0.59

ific

0.56

ja

0.55

 Racing

0.55

 Chess

0.54

ifer

0.54

mar

0.54

 residual

0.53

 (Â£

0.53

Activations Density 0.224%

information related to political figures and discussions about political strategies and alliances

mentions of government officials and positions

No Comments

No Known Activations