INDEX

Explanations

instances where there is conflict or disagreement

oai_token-act-pair · gpt-3.5-turbo

instances of contrast or disagreement in a context of discussion or debate

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

ahime

-0.68

tnc

-0.68

oir

-0.66

 Become

-0.66

pmwiki

-0.66

RIP

-0.66

shift

-0.65

fuck

-0.64

oak

-0.64

mes

-0.64

POSITIVE LOGITS

 cautioned

1.51

 insisted

1.46

 conceded

1.38

 concedes

1.32

 stressed

1.30

 warned

1.26

 insists

1.23

 acknowledged

1.23

 admits

1.22

 emphasized

1.22

Activations Density 0.131%

instances where there is conflict or disagreement

instances of contrast or disagreement in a context of discussion or debate

No Comments

No Known Activations