INDEX

Explanations

references to the concept of "cease", such as "ceasefire" or "cease and desist"

oai_token-act-pair · gpt-3.5-turbo

references to cease or ceasefire agreements

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

ramid

-0.80

swick

-0.79

assium

-0.79

 Inquisitor

-0.78

OTE

-0.75

rium

-0.74

é¾įå

-0.74

Ranked

-0.74

uries

-0.72

VERTISEMENT

-0.71

POSITIVE LOGITS

 cease

0.92

ĸļ

0.88

rences

0.80

²¾

0.79

 seiz

0.76

 ceases

0.73

fires

0.73

lessly

0.72

plin

0.71

cedes

0.70

Activations Density 0.009%

references to the concept of "cease", such as "ceasefire" or "cease and desist"

references to cease or ceasefire agreements

No Comments

No Known Activations