INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

APD

-0.70

 Argon

-0.65

 Nadu

-0.64

vik

-0.63

\\\\\\\\

-0.62

 Riverside

-0.61

ngth

-0.61

 Huntington

-0.60

amb

-0.59

McH

-0.59

POSITIVE LOGITS

 guiActiveUnfocused

0.99

mort

0.76

vez

0.72

 srfAttach

0.72

pour

0.71

Reviewer

0.68

analysis

0.67

tale

0.65

 borne

0.64

enough

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.