INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uador

-0.83

rouse

-0.75

 Rico

-0.72

 ILCS

-0.69

owler

-0.66

?)

-0.65

hack

-0.64

vas

-0.64

 Upton

-0.63

 Bolton

-0.63

POSITIVE LOGITS

noon

0.70

ãĤ°

0.65

aton

0.65

photos

0.64

ã

0.61

acion

0.60

da

0.60

rite

0.60

 Celestial

0.59

âĺ

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.