INDEX

Explanations

words related to opposition, challenge, or negation

oai_token-act-pair · gpt-3.5-turbo

references to counterproductive actions or strategies

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Known

-0.77

 Forge

-0.76

doms

-0.72

 Roche

-0.70

 Thumbnails

-0.67

ABE

-0.65

OOD

-0.64

 Leaving

-0.64

 Tycoon

-0.64

 Flesh

-0.63

POSITIVE LOGITS

counter

1.17

measures

1.08

ror

1.07

attack

1.06

intuitive

1.04

intelligence

1.03

balance

1.03

balanced

0.97

offensive

0.95

rup

0.93

Activations Density 0.009%

words related to opposition, challenge, or negation

references to counterproductive actions or strategies

No Comments

No Known Activations