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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.81

 Dmit

-0.77

ãĥ³ãĤ¸

-0.73

©¶æ

-0.72

è£ħ

-0.71

 Mehran

-0.68

ãĥĥ

-0.68

 Tsarnaev

-0.68

ATIVE

-0.68

ËĪ

-0.67

POSITIVE LOGITS

 execution

0.70

 track

0.69

 stripes

0.65

 gateway

0.64

 graves

0.64

cas

0.63

 digging

0.63

 detectors

0.62

 casing

0.62

 compan

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.