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

 preval

-0.72

 obser

-0.71

 bystand

-0.69

 Veter

-0.68

 Merlin

-0.65

ãĤ¦ãĤ¹

-0.65

mort

-0.64

 Morse

-0.64

 ingred

-0.64

 patented

-0.63

POSITIVE LOGITS

OHN

0.76

SourceFile

0.73

ooter

0.71

ECK

0.68

[_

0.66

CHAT

0.66

é¾į

0.66

RAFT

0.65

Termin

0.65

JJ

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.