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

isSpecialOrderable

-0.82

merce

-0.77

DAQ

-0.74

ramid

-0.74

natureconservancy

-0.74

rimination

-0.72

bnb

-0.72

chance

-0.69

ioxide

-0.68

è¦ļéĨĴ

-0.68

POSITIVE LOGITS

mad

0.82

 Written

0.71

elin

0.66

alter

0.65

kins

0.65

ENT

0.64

initely

0.64

runners

0.62

EXT

0.61

ender

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.