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

zh

-0.76

 Detected

-0.74

Fs

-0.73

 Blues

-0.70

hs

-0.70

TE

-0.68

 Formation

-0.67

scope

-0.67

 Horizon

-0.67

RI

-0.64

POSITIVE LOGITS

¿½

0.79

alling

0.78

oubted

0.73

alty

0.73

interstitial

0.72

fecture

0.71

ificant

0.69

culated

0.69

ition

0.69

omnia

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.