INDEX

Explanations

phrases related to legal and criminal activities, including actions, allegations, and accusations

oai_token-act-pair · gpt-3.5-turbo

references to legal actions and allegations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 ç¥ŀ

-0.87

}}}

-0.83

enment

-0.79

ankind

-0.76

iasm

-0.75

Americ

-0.71

!.

-0.68

Universal

-0.67

 Debate

-0.67

universal

-0.66

POSITIVE LOGITS

 unlawfully

0.82

 inappropriately

0.80

 improper

0.79

 unlawful

0.79

 inappropriate

0.79

 detectives

0.78

 Defendant

0.74

 orally

0.74

 burglary

0.73

 incrim

0.73

Activations Density 0.571%

phrases related to legal and criminal activities, including actions, allegations, and accusations

references to legal actions and allegations

No Comments

No Known Activations