INDEX

Explanations

news-related terms and organizations

oai_token-act-pair · gpt-3.5-turbo

phrases that reference police reports or incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

abiding

-0.65

 enjoys

-0.63

ãĥİ

-0.61

 Tradable

-0.58

Mods

-0.57

yss

-0.57

 predomin

-0.56

atible

-0.56

was

-0.55

pox

-0.54

POSITIVE LOGITS

0.86

.ï¿½

0.82

reports

0.79

.[

0.77

 spokeswoman

0.77

 spokesman

0.77

>.

0.76

 estimates

0.76

.''

0.75

.'

0.75

Activations Density 0.286%

news-related terms and organizations

phrases that reference police reports or incidents

No Comments

No Known Activations