INDEX

Explanations

names of people or entities within a context

oai_token-act-pair · gpt-3.5-turbo

mentions of names in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

nuts

-0.70

ocity

-0.67

nw

-0.67

issance

-0.67

course

-0.66

Pros

-0.65

fps

-0.65

 Instruction

-0.64

pheus

-0.63

rational

-0.63

POSITIVE LOGITS

 redacted

1.27

 withheld

1.03

 blurred

0.97

 tattoo

0.92

 censored

0.89

 engraved

0.87

 obscured

0.85

 etched

0.83

 synonymous

0.83

 pseudonym

0.80

Activations Density 0.155%

names of people or entities within a context

mentions of names in various contexts

No Comments

No Known Activations