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

 Caesar

-0.77

knife

-0.68

 Baal

-0.67

Els

-0.65

yrs

-0.63

-0.61

tv

-0.61

 Sovereign

-0.61

iership

-0.61

 seasoning

-0.60

POSITIVE LOGITS

 Schwar

0.81

 Kash

0.71

ATIONS

0.70

 Shap

0.69

ument

0.68

uve

0.67

ansk

0.65

identally

0.65

ations

0.64

 Sind

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.