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

 pronoun

-0.65

 fate

-0.64

 chars

-0.63

 RIGHT

-0.63

 caster

-0.63

 amulet

-0.62

 ling

-0.62

 orcs

-0.61

 voice

-0.58

 length

-0.58

POSITIVE LOGITS

Sense

0.77

Unity

0.77

Urban

0.76

glomer

0.76

Hack

0.75

Hen

0.75

Toy

0.74

wagen

0.74

ª

0.71

puting

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.