INDEX

Explanations

references to death or dying

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "die" and its context

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

Altern

-0.71

orney

-0.69

artney

-0.66

 Mand

-0.65

rouse

-0.65

OPER

-0.64

OR

-0.63

Mand

-0.63

Kag

-0.62

Grab

-0.62

POSITIVE LOGITS

die

1.09

getic

0.90

 dies

0.84

bold

0.82

fighter

0.81

ffen

0.78

die

0.76

fighters

0.74

 dying

0.74

ously

0.73

Activations Density 0.008%

references to death or dying

occurrences of the word "die" and its context

No Comments

No Known Activations