INDEX

Explanations

text snippets related to spelling words out

oai_token-act-pair · gpt-3.5-turbo

references to the act of spelling

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

romy

-0.76

elaide

-0.75

ribut

-0.74

pheus

-0.74

ivals

-0.73

Kin

-0.71

Reviewer

-0.70

RESULTS

-0.68

 Greenwald

-0.67

 Belg

-0.67

POSITIVE LOGITS

 spelling

1.27

 spelled

1.12

 spell

0.89

bee

0.89

 phon

0.84

 abbrevi

0.83

 shortened

0.80

 pronunciation

0.79

 deaf

0.78

casting

0.78

Activations Density 0.014%

text snippets related to spelling words out

references to the act of spelling

No Comments

No Known Activations