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

 halftime

-0.77

 Tigers

-0.75

\)

-0.71

Cooldown

-0.71

;;;;;;;;;;;;

-0.70

externalActionCode

-0.70

\",

-0.69

Footnote

-0.69

],[

-0.68

sbm

-0.68

POSITIVE LOGITS

master

0.84

ardo

0.82

rency

0.81

iven

0.79

hari

0.77

gan

0.76

gart

0.76

pheus

0.75

aid

0.74

atto

0.74

Activations Density 0.000%

No Known Activations

This feature has no known activations.