INDEX

Explanations

phrases related to disagreement or contradiction

oai_token-act-pair · gpt-3.5-turbo

negative statements and expressions of disagreement

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

 guiActiveUnfocused

-0.94

ãĥĻ

-0.82

resents

-0.74

ļéĨĴ

-0.72

chnology

-0.71

escription

-0.70

skirts

-0.69

edom

-0.66

oxide

-0.65

CAST

-0.65

POSITIVE LOGITS

 however

1.49

 though

1.29

 alas

1.08

 moreover

0.99

 incidentally

0.97

 according

0.96

 meanwhile

0.95

 says

0.95

 anyway

0.88

too

0.85

Activations Density 0.286%

phrases related to disagreement or contradiction

negative statements and expressions of disagreement

No Comments

No Known Activations