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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 heartbeat

-0.66

ropy

-0.60

 Loch

-0.59

 thirds

-0.58

icals

-0.58

 Nano

-0.58

 stink

-0.58

ivas

-0.57

href

-0.56

hops

-0.56

POSITIVE LOGITS

CLAIM

0.77

esta

0.72

è¦ļéĨĴ

0.70

oux

0.65

rical

0.64

 Territories

0.64

quire

0.63

 Upload

0.63

aze

0.63

uct

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.