INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

acio

-0.69

ense

-0.66

Iv

-0.62

 airs

-0.61

rid

-0.61

htt

-0.60

nat

-0.60

))))

-0.60

idd

-0.59

ateg

-0.57

POSITIVE LOGITS

rome

0.75

roma

0.74

bery

0.73

ochet

0.71

roo

0.70

Beast

0.68

ql

0.66

bryce

0.66

devices

0.65

ĪĴ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.