INDEX

Explanations

text related to physical scars

oai_token-act-pair · gpt-3.5-turbo

references to scars and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ablishment

-0.83

eger

-0.75

gaard

-0.74

ullivan

-0.73

hower

-0.71

geist

-0.70

£ı

-0.69

 GEAR

-0.66

aminer

-0.65

ostics

-0.63

POSITIVE LOGITS

 scar

1.16

 scars

1.01

red

0.91

ring

0.91

crow

0.88

lets

0.88

Scar

0.83

fing

0.81

bed

0.79

let

0.78

Activations Density 0.005%

text related to physical scars

references to scars and related terms

No Comments

No Known Activations