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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

thia

-0.70

REM

-0.70

bda

-0.69

-0.68

amera

-0.66

aning

-0.66

azo

-0.66

VICE

-0.65

ocument

-0.63

-0.61

POSITIVE LOGITS

 Immunity

0.83

 Trib

0.74

 Scientists

0.70

''.

0.66

nir

0.63

 Sisters

0.63

 Programs

0.62

bench

0.62

},"

0.61

 quarters

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.