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

 Pixie

-0.79

coat

-0.76

ãĥĩãĤ£

-0.76

ãĤ§

-0.74

 Hail

-0.74

ãĥ³ãĤ¸

-0.72

lance

-0.71

 Pryor

-0.71

gerald

-0.70

ãĥ£

-0.69

POSITIVE LOGITS

Thumbnail

0.79

 invari

0.65

omen

0.65

 elevated

0.61

 amen

0.60

 usher

0.59

 unpre

0.58

 artisan

0.58

 authenticated

0.58

unal

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.