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

heid

-0.73

olitan

-0.73

cup

-0.69

 drink

-0.68

 impeachment

-0.67

 soda

-0.66

 truce

-0.63

rine

-0.63

 Liqu

-0.62

 boycott

-0.61

POSITIVE LOGITS

planet

0.78

tal

0.73

estate

0.73

enhagen

0.72

humans

0.72

aretz

0.72

0.71

detail

0.70

hs

0.67

gnu

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.