INDEX

Explanations

phrases indicating suffering or harm inflicted on individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "suffered" in various contexts related to hardship or adversity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 bloc

-0.75

rouse

-0.73

ranean

-0.70

azi

-0.66

send

-0.64

cle

-0.64

urry

-0.64

esters

-0.63

atures

-0.63

ococ

-0.63

POSITIVE LOGITS

 suffered

1.09

 setbacks

0.87

Downloadha

0.85

ãĤ´ãĥ³

0.84

lehem

0.82

 inflicted

0.81

 suffers

0.79

 losses

0.78

 horrend

0.78

 Loss

0.78

Activations Density 0.009%

phrases indicating suffering or harm inflicted on individuals

instances of the word "suffered" in various contexts related to hardship or adversity

No Comments

No Known Activations