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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

renheit

-0.92

nels

-0.90

iversary

-0.83

stories

-0.80

uries

-0.80

Tea

-0.79

wright

-0.77

thood

-0.77

 Eliot

-0.77

ivities

-0.75

POSITIVE LOGITS

orp

0.79

 redes

0.79

 shorth

0.71

 analges

0.69

 herpes

0.65

PAL

0.64

 bilateral

0.64

 indicated

0.64

 Nordic

0.62

 Tasman

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.