INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

pton

-0.71

 Miliband

-0.65

âĹ¼

-0.63

 alcoholic

-0.60

breaker

-0.60

Ob

-0.58

oper

-0.58

 Coffin

-0.58

MED

-0.58

Tes

-0.57

POSITIVE LOGITS

ĪĴ

0.82

nces

0.79

earch

0.77

 discrep

0.75

afety

0.74

Ü

0.73

 tremend

0.73

eval

0.72

sit

0.71

Ļ½

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.