INDEX

Explanations

specific mentions or instances of an opponent

oai_token-act-pair · gpt-3.5-turbo

mentions of an opponent across various contexts

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

mberg

-0.76

redits

-0.75

olia

-0.71

angering

-0.70

oots

-0.68

ening

-0.66

vest

-0.66

urgical

-0.65

onics

-0.64

acca

-0.62

POSITIVE LOGITS

 opponent

1.28

 opponents

1.01

foe

1.00

onent

0.92

 adversary

0.83

 batters

0.80

 combatants

0.79

 challenger

0.78

 strugg

0.77

 defeats

0.73

Activations Density 0.011%

specific mentions or instances of an opponent

mentions of an opponent across various contexts

No Comments

No Known Activations