INDEX

Explanations

words related to the concept of 'pieces'

oai_token-act-pair · gpt-3.5-turbo

concepts related to themes of fidelity and imitation

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

 Ples

-0.66

 Graves

-0.63

 descended

-0.62

Ur

-0.62

 Scand

-0.62

auc

-0.61

 relocated

-0.61

ãĥİ

-0.60

intestinal

-0.60

lower

-0.60

POSITIVE LOGITS

 Pieces

1.63

 imitation

1.24

idelity

1.09

eday

0.82

 obedience

0.78

ventus

0.76

izabeth

0.74

edience

0.72

 emulation

0.72

 Piece

0.71

Activations Density 0.009%

words related to the concept of 'pieces'

concepts related to themes of fidelity and imitation

No Comments

No Known Activations