INDEX

Explanations

references to seeking or enacting revenge

oai_token-act-pair · gpt-3.5-turbo

terms related to revenge and vengeance

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

 livest

-0.87

uddle

-0.76

ographics

-0.74

anian

-0.67

 Transition

-0.66

hover

-0.64

regulated

-0.64

66666666

-0.63

kered

-0.63

 interchange

-0.61

POSITIVE LOGITS

 against

1.22

 Against

1.05

against

1.05

 retaliation

1.04

 revenge

1.02

 vengeance

0.95

 retali

0.95

 retribution

0.95

Against

0.93

 repr

0.91

Activations Density 0.075%

references to seeking or enacting revenge

terms related to revenge and vengeance

No Comments

No Known Activations