INDEX

Explanations

words related to negative actions or characteristics, specifically focusing on petty behavior

oai_token-act-pair · gpt-3.5-turbo

references to trivial or minor offenses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Downloadha

-0.89

ahead

-0.78

hov

-0.77

ioch

-0.76

hetti

-0.76

ources

-0.76

avez

-0.74

heimer

-0.73

Recomm

-0.73

igslist

-0.72

POSITIVE LOGITS

 petty

0.91

cipled

0.79

 theft

0.77

 arithmetic

0.72

 Petty

0.72

 misdemeanor

0.72

 provocation

0.67

tru

0.66

fel

0.66

 Theft

0.66

Activations Density 0.014%

words related to negative actions or characteristics, specifically focusing on petty behavior

references to trivial or minor offenses

No Comments

No Known Activations