INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

-0.68

REDACTED

-0.65

 Providence

-0.65

tem

-0.64

 Tosh

-0.62

reason

-0.62

 prud

-0.61

 intrins

-0.59

 Incident

-0.58

POSITIVE LOGITS

ommel

0.89

olkien

0.77

bered

0.75

 scrut

0.75

anyl

0.72

earchers

0.72

irie

0.71

 handc

0.71

natureconservancy

0.69

eworld

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.