INDEX

Explanations

mentions of the word "alcoholic", especially pertaining to people

oai_token-act-pair · gpt-3.5-turbo

terms related to addiction and loss of control

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

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

foundation

-0.83

ding

-0.79

arity

-0.74

bley

-0.74

imble

-0.73

artisan

-0.72

ĺħ

-0.70

eker

-0.69

ĪĴ

-0.69

tsky

-0.68

POSITIVE LOGITS

ABLE

0.90

uous

0.83

 beverages

0.76

XVI

0.74

uations

0.73

 involuntary

0.72

 alcoholism

0.71

ãĥ¯ãĥ³

0.70

âĶĢâĶĢ

0.70

URA

0.69

Activations Density 0.023%

mentions of the word "alcoholic", especially pertaining to people

terms related to addiction and loss of control

No Comments

No Known Activations