INDEX

Explanations

occurrences of the word "bloody" highlighting violence or brutality

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

strong, negative emotions or actions

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

PLIED

-0.85

Demand

-0.83

BOOK

-0.82

Reviewer

-0.80

berman

-0.78

anol

-0.76

CHAT

-0.75

Introduced

-0.71

perm

-0.71

Rate

-0.70

POSITIVE LOGITS

 bloody

0.99

 noses

0.87

 slaughter

0.74

stained

0.72

 prick

0.72

 bast

0.71

 stopp

0.70

 swat

0.70

cade

0.70

 wounds

0.70

Activations Density 0.007%

occurrences of the word "bloody" highlighting violence or brutality

strong, negative emotions or actions

No Comments

No Known Activations