INDEX

Explanations

contractions of words ending in 'not'

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "couldn't" and its variations, indicating inability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 blitz

-0.74

 Jinn

-0.71

 humanities

-0.66

 radicals

-0.65

 Julius

-0.64

 vulner

-0.63

 chunks

-0.61

 plank

-0.60

 dehuman

-0.60

 scraps

-0.59

POSITIVE LOGITS

İ

0.91

ï¸ı

0.89

tem

0.84

idate

0.83

_>

0.81

ember

0.80

ivably

0.79

itely

0.79

tre

0.79

emb

0.78

Activations Density 0.079%

contractions of words ending in 'not'

instances of the phrase "couldn't" and its variations, indicating inability

No Comments

No Known Activations