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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.70

ious

-0.69

head

-0.68

ned

-0.67

NING

-0.66

bringer

-0.64

 Sachs

-0.64

song

-0.62

EMENT

-0.62

NY

-0.61

POSITIVE LOGITS

 confir

0.96

pse

0.92

©¶æ

0.85

 Agric

0.82

 artif

0.81

igham

0.80

 psychiat

0.79

 destro

0.78

 conduc

0.76

orum

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.