INDEX

Explanations

word phrases related to news articles and specific events such as accidents, criminal cases, and political statements

oai_token-act-pair · gpt-3.5-turbo

phrases related to significant events or legal matters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

mittedly

-0.76

theless

-0.72

TEXTURE

-0.71

áµ

-0.71

fortunately

-0.70

ÙĴ

-0.69

Newsletter

-0.68

paralle

-0.68

 embodied

-0.66

eday

-0.66

POSITIVE LOGITS

 Kavanaugh

0.87

 Says

0.86

 probe

0.86

 Replay

0.85

 Copyright

0.82

 Dems

0.77

 vigil

0.76

 feds

0.75

 Probe

0.74

 dies

0.73

Activations Density 1.150%

word phrases related to news articles and specific events such as accidents, criminal cases, and political statements

phrases related to significant events or legal matters

No Comments

No Known Activations