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

otos

-0.82

omo

-0.79

ortment

-0.79

apses

-0.78

adish

-0.77

otions

-0.75

ptions

-0.74

esty

-0.73

ugs

-0.72

ocrats

-0.72

POSITIVE LOGITS

Long

0.72

Running

0.70

 long

0.70

 Mobil

0.68

 Assistance

0.66

Coll

0.65

Surv

0.65

AFL

0.63

0.62

 liner

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.