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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Ness

-0.82

athe

-0.78

 Beir

-0.72

chio

-0.71

SS

-0.71

fing

-0.68

Semitism

-0.68

ularity

-0.67

sch

-0.66

TeX

-0.64

POSITIVE LOGITS

 dehyd

0.72

 Ãľ

0.63

RPM

0.62

 humid

0.61

 decriminal

0.60

interstitial

0.60

 snap

0.59

 RESP

0.58

 rust

0.57

 cracked

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.