INDEX

Explanations

phrases with the word "hint"

oai_token-act-pair · gpt-3.5-turbo

phrases indicating subtle hints or implications

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

frey

-0.71

artney

-0.67

nea

-0.67

bred

-0.67

lux

-0.67

ccording

-0.65

cus

-0.65

CENT

-0.64

fare

-0.63

portion

-0.63

POSITIVE LOGITS

 hint

1.52

 hints

1.32

 clue

0.89

 hinted

0.88

 glimps

0.83

 wink

0.82

 warning

0.81

 clues

0.80

 intim

0.72

ibly

0.71

Activations Density 0.021%

phrases with the word "hint"

phrases indicating subtle hints or implications

No Comments

No Known Activations