INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to significant traumatic events or personal crises

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 ourselves

-0.62

urches

-0.61

ept

-0.60

 disclaimer

-0.59

 issu

-0.58

anty

-0.58

asma

-0.57

afety

-0.56

 selves

-0.56

 GOODMAN

-0.55

POSITIVE LOGITS

 himself

1.23

his

1.12

 hers

1.08

his

0.91

 veto

0.86

 Himself

0.85

him

0.81

His

0.78

HIS

0.77

 theirs

0.77

Activations Density 0.959%

proper nouns

references to significant traumatic events or personal crises

No Comments

No Known Activations