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

ĸļ

-0.73

stein

-0.67

icon

-0.67

Ń·

-0.67

skin

-0.66

asso

-0.66

emon

-0.65

ileged

-0.65

 reconc

-0.65

acci

-0.65

POSITIVE LOGITS

Contract

0.84

 Contracts

0.78

 Contract

0.74

Ùĩ

0.72

 Committees

0.67

Published

0.66

 clauses

0.66

ç«

0.65

Recomm

0.65

Ø¹

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.