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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

)].

-0.79

thing

-0.73

ylum

-0.69

 Giul

-0.66

thur

-0.66

blocks

-0.65

minist

-0.64

tackle

-0.63

inline

-0.62

 pharmacies

-0.62

POSITIVE LOGITS

Coll

0.73

¥µ

0.72

Åį

0.69

orsche

0.67

 incarnation

0.67

artz

0.66

¿½

0.63

otal

0.62

hon

0.62

pmwiki

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.