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

Dialogue

-0.80

Radio

-0.75

Journal

-0.75

Relations

-0.72

yip

-0.70

Done

-0.70

Mem

-0.69

 âĢİ

-0.68

 Decision

-0.66

Present

-0.65

POSITIVE LOGITS

 stitches

0.80

coh

0.76

 baptized

0.73

 bart

0.72

 ages

0.71

 overdoses

0.65

 defic

0.65

 stitching

0.65

igr

0.63

bold

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.