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

Not in Any Lists

No Comments

Negative Logits

hess

-0.82

 Indy

-0.70

atter

-0.68

alde

-0.66

phabet

-0.66

serv

-0.64

earth

-0.64

pur

-0.62

uni

-0.61

arna

-0.61

POSITIVE LOGITS

Cth

0.82

 suspic

0.82

ertodd

0.78

 fortun

0.70

 OPEC

0.70

 horizont

0.70

pmwiki

0.70

 neighb

0.69

 Siem

0.66

itars

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.