INDEX

Explanations

phrases related to law enforcement and criminal activities

oai_token-act-pair · gpt-3.5-turbo

references to social class and systemic inequalities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 fortun

-0.72

 sacrific

-0.63

 Bigfoot

-0.63

 flavour

-0.62

Yor

-0.61

 Garmin

-0.61

GGGGGGGG

-0.60

bun

-0.60

 Argon

-0.59

CLR

-0.59

POSITIVE LOGITS

ï¸ı

0.94

baugh

0.81

few

0.79

gypt

0.76

ploy

0.76

lishes

0.75

0.73

Pg

0.73

own

0.73

ately

0.72

Activations Density 0.452%

phrases related to law enforcement and criminal activities

references to social class and systemic inequalities

No Comments

No Known Activations