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

Not in Any Lists

No Comments

Negative Logits

ennes

-0.75

 homeland

-0.71

enic

-0.70

amy

-0.70

 retribution

-0.69

 plummet

-0.67

 curtail

-0.65

 indefinitely

-0.64

retion

-0.64

 bars

-0.63

POSITIVE LOGITS

Spec

0.87

Street

0.84

Ire

0.80

Appearances

0.80

tek

0.78

ISO

0.77

natureconservancy

0.76

é¾įåĸļå£«

0.76

Streamer

0.76

ĪĴ

0.74

Activations Density 0.000%

No Known Activations

This feature has no known activations.