INDEX

Explanations

mentions of personal preferences or favorites

oai_token-act-pair · gpt-3.5-turbo

references to favorite things or preferences

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

redits

-0.73

ulative

-0.73

avis

-0.73

lam

-0.73

aping

-0.71

uid

-0.70

DEF

-0.70

proof

-0.70

usted

-0.70

compliance

-0.69

POSITIVE LOGITS

 haun

1.01

 haunt

0.91

 beverage

0.87

 hobby

0.86

 tunes

0.86

 hobbies

0.86

 snack

0.83

 underdog

0.81

 childhood

0.81

 meal

0.78

Activations Density 0.059%

mentions of personal preferences or favorites

references to favorite things or preferences

No Comments

No Known Activations