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

ylum

-0.72

olean

-0.65

CHR

-0.61

eals

-0.59

 Appears

-0.58

bourg

-0.57

terday

-0.57

eworks

-0.57

MIS

-0.57

 Addiction

-0.57

POSITIVE LOGITS

atche

0.80

tered

0.77

roma

0.76

utenberg

0.73

iago

0.72

nell

0.72

uc

0.67

riger

0.65

clinton

0.65

pass

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.