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

Aren

-0.65

 appell

-0.61

talk

-0.60

 Stall

-0.59

xual

-0.58

 actresses

-0.58

hook

-0.58

plet

-0.58

irlf

-0.58

 trains

-0.58

POSITIVE LOGITS

eneg

0.92

uncture

0.78

ZA

0.73

ayn

0.71

heid

0.71

ICLE

0.70

å§«

0.70

ierre

0.67

zu

0.66

oreal

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.