INDEX

Explanations

various scenarios or states related to negative experiences such as harm, fighting, or being in need

oai_token-act-pair · gpt-3.5-turbo

references to harm, risk, and experiences of suffering

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Fla

-0.58

-0.56

Ala

-0.54

 oranges

-0.54

 Tunis

-0.54

 Ramos

-0.54

 doubtless

-0.53

 Vale

-0.52

 Albania

-0.52

 Columb

-0.52

POSITIVE LOGITS

 thereof

1.20

depending

1.06

abouts

1.05

 versa

0.98

Else

0.93

 altogether

0.92

 otherwise

0.87

omever

0.86

 combination

0.83

 equivalent

0.82

Activations Density 0.457%

various scenarios or states related to negative experiences such as harm, fighting, or being in need

references to harm, risk, and experiences of suffering

No Comments

No Known Activations