INDEX

Explanations

phrases where something is not exactly what it seems or expected to be

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize a sense of precision or accuracy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

idas

-0.82

iewicz

-0.81

otation

-0.80

tailed

-0.76

çļ

-0.74

soon

-0.74

isson

-0.73

aughs

-0.73

ð

-0.72

ription

-0.72

POSITIVE LOGITS

 conducive

0.87

 pleasant

0.75

 fool

0.75

 rocket

0.74

 thrilled

0.72

 unbiased

0.72

 reassuring

0.72

 stellar

0.71

 forgiving

0.70

shy

0.69

Activations Density 0.042%

phrases where something is not exactly what it seems or expected to be

phrases that emphasize a sense of precision or accuracy

No Comments

No Known Activations