INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

é¾į

-0.85

iquid

-0.80

é¾įå¥ĳå£«

-0.79

itect

-0.79

ailability

-0.78

ufact

-0.78

ãĥ´ãĤ¡

-0.78

ãĤ·ãĥ£

-0.78

ãĤ¢ãĥ«

-0.77

dimension

-0.76

POSITIVE LOGITS

 Castle

0.66

 detention

0.64

 Yard

0.63

ater

0.62

 Citation

0.61

 Lash

0.61

aters

0.59

Wim

0.59

lege

0.58

Wad

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.