INDEX

Explanations

words related to curses

oai_token-act-pair · gpt-3.5-turbo

terms related to curses, misfortunes, or negative consequences

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

å§«

-0.84

olitan

-0.83

nington

-0.81

visors

-0.79

arger

-0.78

galitarian

-0.76

utm

-0.75

unicip

-0.75

·

-0.75

romy

-0.72

POSITIVE LOGITS

 curse

1.18

 curses

1.04

 cursed

0.92

 cursing

0.88

 Curse

0.77

 cringe

0.75

 lashes

0.71

hammer

0.69

 destro

0.69

fully

0.68

Activations Density 0.007%

words related to curses

terms related to curses, misfortunes, or negative consequences

No Comments

No Known Activations