INDEX

Explanations

mentions or discussions about archives

oai_token-act-pair · gpt-3.5-turbo

terms related to archives

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

rollers

-0.78

edes

-0.66

inton

-0.63

 perceived

-0.61

Fit

-0.59

 Crow

-0.59

Ã®

-0.59

 pronounced

-0.59

ievers

-0.58

å¥

-0.57

POSITIVE LOGITS

 archive

3.83

 archives

3.04

 Archive

2.47

archive

2.22

 Archives

2.13

 archived

2.01

 catalogue

1.38

 arch

1.37

Arch

1.30

archives

1.28

Activations Density 0.018%

mentions or discussions about archives

terms related to archives

No Comments

No Known Activations