INDEX

Explanations

terms or phrases related to official approval processes

oai_token-act-pair · gpt-3.5-turbo

terms related to legal and regulatory approvals

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

 metaphors

-0.69

 lately

-0.69

 misunderstand

-0.66

 prank

-0.66

 chatter

-0.65

 distractions

-0.64

 nutshell

-0.64

 shortcuts

-0.63

 overshadowed

-0.62

 speculate

-0.61

POSITIVE LOGITS

 certify

1.11

 cert

0.97

cert

0.94

 certification

0.93

 certified

0.92

 beforehand

0.90

 preferably

0.88

 whichever

0.88

worthiness

0.87

 approval

0.87

Activations Density 0.435%

terms or phrases related to official approval processes

terms related to legal and regulatory approvals

No Comments

No Known Activations