INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ACK

-0.68

acks

-0.67

FINE

-0.64

ileged

-0.64

everal

-0.63

airs

-0.62

 sqor

-0.61

existence

-0.60

 RELE

-0.59

PUT

-0.58

POSITIVE LOGITS

 Swansea

0.71

 Bale

0.70

opia

0.68

reon

0.68

 SpaceX

0.68

Naz

0.67

Kro

0.65

sburg

0.65

 Warwick

0.64

 Cory

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.