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

ļéĨĴ

-0.79

 aback

-0.76

SELECT

-0.74

staking

-0.74

 contrace

-0.71

culosis

-0.71

 stranded

-0.70

 entangled

-0.70

 ineligible

-0.69

 cache

-0.68

POSITIVE LOGITS

 Frazier

0.69

rieg

0.68

essors

0.68

dq

0.67

amen

0.66

rolet

0.65

atus

0.64

amic

0.64

fx

0.62

 grit

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.