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

jab

-0.71

utable

-0.71

ropolitan

-0.70

utsu

-0.67

onga

-0.67

ariat

-0.66

anche

-0.66

 Kinnikuman

-0.65

iked

-0.65

 bulldo

-0.64

POSITIVE LOGITS

SSL

0.76

BuyableInstoreAndOnline

0.71

rine

0.71

inet

0.67

ns

0.66

yards

0.65

STD

0.65

Shin

0.65

pid

0.65

Ident

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.