INDEX

Explanations

proper nouns and phrases related to individuals, organizations, and events mentioned in news articles

oai_token-act-pair · gpt-3.5-turbo

references to authoritative sources or statements regarding events or decisions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 miscarriage

-0.53

 miracle

-0.51

abiding

-0.50

pires

-0.50

 enjoys

-0.50

successfully

-0.49

 domest

-0.49

 harms

-0.49

 diaper

-0.47

 paternity

-0.47

POSITIVE LOGITS

 quoted

0.61

ipop

0.58

zinski

0.56

 adding

0.56

 spokesman

0.56

 Assistant

0.55

¶æ

0.55

 spokeswoman

0.54

iliary

0.53

ibrarian

0.53

Activations Density 0.870%

proper nouns and phrases related to individuals, organizations, and events mentioned in news articles

references to authoritative sources or statements regarding events or decisions

No Comments

No Known Activations