INDEX

Explanations

terms related to physical wounds

oai_token-act-pair · gpt-3.5-turbo

references to wounds or injuries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

*/(

-0.83

alty

-0.79

poons

-0.77

anooga

-0.75

effic

-0.69

ĨĴ

-0.66

ordan

-0.64

 gravity

-0.64

initions

-0.63

ourke

-0.63

POSITIVE LOGITS

robe

0.84

 wounds

0.84

burn

0.82

 wound

0.82

strap

0.81

terson

0.77

bow

0.76

 Whedon

0.74

bolt

0.70

stead

0.70

Activations Density 0.014%

terms related to physical wounds

references to wounds or injuries

No Comments

No Known Activations