INDEX

Explanations

mentions of sports players, specifically focusing on midfielders

oai_token-act-pair · gpt-3.5-turbo

references to specific soccer player positions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aber

-0.74

igslist

-0.74

ãĥ¼ãĥ³

-0.71

rip

-0.71

road

-0.66

bart

-0.65

Lear

-0.64

ories

-0.61

trak

-0.60

TIME

-0.59

POSITIVE LOGITS

 midfielder

0.86

 winger

0.81

 extraord

0.78

 footballer

0.74

 defender

0.73

 Liam

0.70

 Marcel

0.70

duo

0.70

 fielder

0.69

 outfielder

0.67

Activations Density 0.015%

mentions of sports players, specifically focusing on midfielders

references to specific soccer player positions

No Comments

No Known Activations