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

Negative Logits

psons

-0.76

 Seym

-0.67

 protester

-0.64

 Czech

-0.63

 Dispatch

-0.62

 Sunderland

-0.62

 Bullets

-0.62

 Trace

-0.60

 sprayed

-0.59

 Hussain

-0.59

POSITIVE LOGITS

ngth

0.91

iary

0.79

space

0.79

ËĪ

0.77

lishes

0.70

issance

0.69

ned

0.68

dfx

0.68

fee

0.67

vu

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.