INDEX

Explanations

mentions of specified letter combinations, possibly related to a code or specific language use

oai_token-act-pair · gpt-3.5-turbo

instances where large sums of money or significant offers are mentioned

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 citiz

-0.73

 ageing

-0.67

 tremend

-0.66

etheless

-0.65

 nutshell

-0.65

 nomine

-0.64

minist

-0.64

 footing

-0.63

 endeav

-0.62

 wiser

-0.62

POSITIVE LOGITS

PHOTOS

0.89

Indeed

0.88

Attempts

0.86

Correct

0.80

 Related

0.77

Letter

0.76

Asked

0.75

Chel

0.74

Shape

0.74

Others

0.74

Activations Density 0.261%

mentions of specified letter combinations, possibly related to a code or specific language use

instances where large sums of money or significant offers are mentioned

No Comments

No Known Activations