INDEX

Explanations

phrases indicating the start of a spoken statement or presentation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Speaking"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ery

-0.78

eer

-0.74

ichen

-0.66

eers

-0.65

involved

-0.65

hews

-0.65

eries

-0.62

isd

-0.61

vier

-0.61

feeding

-0.59

POSITIVE LOGITS

 anecd

0.99

 bluntly

0.85

 frankly

0.85

 anonymously

0.83

 candid

0.83

 honestly

0.82

 engagements

0.80

 exclusively

0.79

 personally

0.76

 aloud

0.74

Activations Density 0.025%

phrases indicating the start of a spoken statement or presentation

instances of the word "Speaking"

No Comments

No Known Activations