INDEX

Explanations

mentions of hitting something, whether it's a physical object or a metaphorical target

oai_token-act-pair · gpt-3.5-turbo

instances of the word "hit."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

pires

-0.73

inent

-0.65

agin

-0.65

otype

-0.64

æĢ

-0.63

orld

-0.63

ç«

-0.59

UTH

-0.57

aer

-0.56

udes

-0.56

POSITIVE LOGITS

ched

1.19

boxes

0.96

ches

0.92

 puberty

0.89

 hardest

0.79

box

0.77

tle

0.76

 milestones

0.76

chens

0.76

peak

0.75

Activations Density 0.033%

mentions of hitting something, whether it's a physical object or a metaphorical target

instances of the word "hit."

No Comments

No Known Activations