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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ibrary

-0.80

alsh

-0.78

chy

-0.71

aukee

-0.70

£ı

-0.65

illard

-0.62

 transcript

-0.62

 hinge

-0.61

 Till

-0.60

 hinges

-0.60

POSITIVE LOGITS

rament

0.79

ofi

0.74

aldi

0.73

Parameter

0.65

olate

0.64

 nano

0.64

Micro

0.63

 Nasa

0.62

ificial

0.61

ppo

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.