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

beat

-0.70

 valuation

-0.68

 Bake

-0.64

housing

-0.62

ixels

-0.62

 messenger

-0.61

wit

-0.60

uries

-0.59

 Seeking

-0.58

Money

-0.58

POSITIVE LOGITS

ļéĨĴ

0.99

aucuses

0.86

âĵĺ

0.82

CLOSE

0.81

catentry

0.80

osponsors

0.76

EngineDebug

0.75

ulhu

0.75

 submar

0.73

jad

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.