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

 Cook

-0.71

 Volunteers

-0.70

 Chow

-0.70

racuse

-0.68

 Grizz

-0.64

 cook

-0.63

REP

-0.63

 Prospect

-0.61

 Lyme

-0.60

å§«

-0.59

POSITIVE LOGITS

pees

0.78

©¶æ¥µ

0.78

dfx

0.77

olith

0.74

HCR

0.72

acly

0.71

oldemort

0.70

gery

0.70

iar

0.68

umat

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.