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

TABLE

-0.66

ornia

-0.65

 brig

-0.64

united

-0.64

 stronghold

-0.64

utenant

-0.64

 labour

-0.62

 biologist

-0.61

iless

-0.60

 Lieutenant

-0.59

POSITIVE LOGITS

uyomi

0.91

ãĤ¦ãĤ¹

0.84

angan

0.78

ãĤ§

0.77

ãĥ¯ãĥ³

0.73

ãĤ£

0.71

APS

0.70

Closure

0.70

Dest

0.69

Buff

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.