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

HI

-0.75

SSL

-0.71

 probation

-0.64

SHIP

-0.63

umblr

-0.61

ESCO

-0.61

CLASSIFIED

-0.60

RIP

-0.59

 triv

-0.58

MSN

-0.58

POSITIVE LOGITS

iencies

0.84

nexus

0.76

romeda

0.75

luence

0.66

lag

0.65

gran

0.64

 Lumpur

0.63

entimes

0.63

ogun

0.62

 kits

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.