INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ype

-0.71

TY

-0.70

stakes

-0.68

 Staten

-0.67

anan

-0.63

Crew

-0.63

ELD

-0.62

REM

-0.61

aptic

-0.61

Args

-0.59

POSITIVE LOGITS

 Tempest

0.61

 Enix

0.61

 dwarves

0.60

romeda

0.60

 Alban

0.60

 Merlin

0.59

lin

0.59

ppo

0.59

elight

0.59

erer

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.