INDEX

Explanations

political figures or media outlets being criticized

oai_token-act-pair · gpt-3.5-turbo

instances of the word "criticized" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

OTE

-0.73

ëĭ

-0.71

bourne

-0.69

mop

-0.68

ther

-0.67

ulhu

-0.66

aho

-0.66

ammy

-0.65

nown

-0.65

mad

-0.64

POSITIVE LOGITS

imaru

0.73

 Cosponsors

0.73

 harshly

0.69

 comments

0.63

 remarks

0.63

 Stab

0.62

him

0.61

Orb

0.61

 critiques

0.61

 critic

0.61

Activations Density 0.050%

political figures or media outlets being criticized

instances of the word "criticized" and its variations

No Comments

No Known Activations