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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

chwitz

-0.84

iers

-0.78

ija

-0.72

inki

-0.70

eu

-0.69

ŃĶ

-0.67

eez

-0.65

gebra

-0.64

rict

-0.63

ghan

-0.61

POSITIVE LOGITS

 Corpus

0.88

 thumbnail

0.73

 Uriel

0.70

 Payton

0.70

REDACTED

0.67

 Pixie

0.66

essors

0.62

href

0.62

 Paramount

0.62

 ashore

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.