INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ross

-0.70

inois

-0.70

ç·

-0.65

rosso

-0.65

idis

-0.64

ãĥĻ

-0.64

onen

-0.61

 Prometheus

-0.60

olulu

-0.60

\":

-0.59

POSITIVE LOGITS

ELD

0.80

ModLoader

0.77

tesy

0.69

 Entered

0.68

theless

0.64

 offic

0.61

ctors

0.60

ges

0.60

vantage

0.59

ingly

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.