INDEX

Explanations

references to gendered terms, particularly associated with males and their behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

 Cosponsors

-0.87

Decre

-0.75

Folder

-0.74

utions

-0.73

pring

-0.72

idate

-0.71

fman

-0.71

atories

-0.70

mberg

-0.69

letcher

-0.68

POSITIVE LOGITS

holes

0.96

hole

0.90

 buddy

0.86

 buddies

0.82

 jeans

0.79

 haircut

0.72

ards

0.72

 dude

0.71

 wink

0.68

who

0.66

Activations Density 0.013%

references to gendered terms, particularly associated with males and their behaviors

No Comments

No Known Activations