INDEX

Explanations

mentions of winning, losing, outcomes, and results

oai_token-act-pair · gpt-3.5-turbo

references to competitive outcomes, participants, and winning/losing scenarios

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

annels

-0.93

dayName

-0.90

DragonMagazine

-0.84

agents

-0.83

 puberty

-0.76

 intrusive

-0.75

urnal

-0.72

 Anthrop

-0.72

 prescribing

-0.72

 psychiatry

-0.70

POSITIVE LOGITS

win

1.87

 wins

1.85

 victory

1.82

 winning

1.70

 loser

1.62

 victories

1.60

 Winning

1.58

 defeat

1.55

 winner

1.54

 defeats

1.52

Activations Density 0.561%

mentions of winning, losing, outcomes, and results

references to competitive outcomes, participants, and winning/losing scenarios

No Comments

No Known Activations