INDEX

Explanations

statements negating or denying some kind of association or influence

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize lack or negation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Eid

-0.66

 Foley

-0.65

 Metatron

-0.65

 Polk

-0.63

 Falk

-0.63

 Citizens

-0.63

 McCabe

-0.61

 Greenberg

-0.61

 Extensions

-0.61

Flo

-0.59

POSITIVE LOGITS

 uncertain

1.04

way

0.99

 hurry

0.96

 doubt

0.89

WAY

0.87

 particular

0.84

osaurs

0.81

xus

0.81

cific

0.81

 danger

0.80

Activations Density 0.036%

statements negating or denying some kind of association or influence

phrases that emphasize lack or negation

No Comments

No Known Activations