INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 aval

-0.68

zn

-0.64

 pleasing

-0.64

 Dane

-0.62

 open

-0.61

ple

-0.59

{*

-0.59

 dele

-0.59

 buoy

-0.59

 user

-0.58

POSITIVE LOGITS

ciating

0.80

anship

0.78

agascar

0.75

ospons

0.74

billion

0.73

SourceFile

0.72

 Cosponsors

0.72

ileged

0.72

rified

0.71

ecause

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.