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

sbm

-0.77

ãĥĥ

-0.76

ãĤ¨ãĥ«

-0.72

hoe

-0.71

oux

-0.69

ritis

-0.65

ird

-0.64

 Maul

-0.64

Ik

-0.64

govtrack

-0.63

POSITIVE LOGITS

 gaming

0.75

Gaming

0.68

 Gaming

0.68

 disemb

0.64

 Electronic

0.64

Ultimate

0.64

gaming

0.60

Reviewer

0.60

Tea

0.57

 Barcelona

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.