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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

resa

-0.65

 DRAG

-0.61

 Unified

-0.60

 Pend

-0.60

 Casting

-0.59

fecture

-0.57

 âĢº

-0.57

 DISTRICT

-0.57

SQU

-0.56

OP

-0.56

POSITIVE LOGITS

 which

1.04

which

0.75

ftime

0.68

ICLE

0.68

claim

0.64

ibaba

0.64

 attribution

0.61

 ourselves

0.60

ikhail

0.60

 something

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.