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

ophon

-0.75

ebin

-0.74

utenberg

-0.74

imes

-0.72

beer

-0.69

Events

-0.67

acs

-0.65

berman

-0.65

Joy

-0.64

_>

-0.64

POSITIVE LOGITS

ioch

0.79

ENA

0.76

ETH

0.75

âĸĳ

0.70

ãĤ·ãĥ£

0.69

-+-+-+-+

0.68

ãĥ´

0.68

ICLE

0.68

VM

0.66

âķĲ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.