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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zai

-0.80

wcs

-0.73

ymm

-0.72

 ACTIONS

-0.72

iage

-0.70

gencies

-0.68

wagen

-0.68

yss

-0.67

ibling

-0.67

idas

-0.66

POSITIVE LOGITS

 perf

0.62

bro

0.59

 listener

0.59

 algorith

0.59

Minor

0.59

 Wallet

0.59

 unbiased

0.58

 Mayo

0.57

 artif

0.57

hemat

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.