INDEX

Explanations

negative terms or insults

oai_token-act-pair · gpt-3.5-turbo

derogatory terms and insults directed at individuals or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

tnc

-0.76

ondo

-0.74

RH

-0.74

ira

-0.73

winner

-0.72

Recomm

-0.70

APH

-0.69

ranging

-0.69

ANC

-0.68

iture

-0.68

POSITIVE LOGITS

 idiots

1.03

 idiot

1.02

 bastard

0.98

 bully

0.97

 spew

0.95

 hypoc

0.91

 bitch

0.90

 asshole

0.89

 sucker

0.88

 crap

0.86

Activations Density 0.072%

negative terms or insults

derogatory terms and insults directed at individuals or groups

No Comments

No Known Activations