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

 fracturing

-0.76

uana

-0.69

 paused

-0.66

EVA

-0.65

 inver

-0.65

staking

-0.63

 isol

-0.63

 setbacks

-0.62

cock

-0.62

 quarantine

-0.62

POSITIVE LOGITS

Kah

0.73

dfx

0.71

Khe

0.71

reme

0.70

Sok

0.70

 Riot

0.70

esp

0.68

idas

0.68

ELS

0.67

asar

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.