INDEX

Explanations

proper nouns consisting of the word "ran"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "ran"

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

GMT

-0.65

lda

-0.63

earable

-0.62

ready

-0.62

£ı

-0.60

¿½

-0.59

 subp

-0.58

cember

-0.57

 chromosome

-0.56

ORK

-0.56

POSITIVE LOGITS

vier

1.06

igans

0.94

fo

0.88

quez

0.86

kees

0.86

egal

0.82

igan

0.82

tenance

0.80

iety

0.79

ussia

0.78

Activations Density 0.018%

proper nouns consisting of the word "ran"

occurrences of the substring "ran"

No Comments

No Known Activations