INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

sburgh

-0.69

 Disabled

-0.67

otor

-0.67

alloc

-0.67

pun

-0.65

puter

-0.65

mares

-0.65

istar

-0.62

borgh

-0.62

å£«

-0.62

POSITIVE LOGITS

aram

0.70

 balcon

0.70

 Geral

0.64

.''.

0.63

Weiss

0.63

Ïģ

0.63

ndum

0.63

 rapp

0.62

rine

0.62

 caut

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.