INDEX

Explanations

proper names or locations in a news context

oai_token-act-pair · gpt-3.5-turbo

recurring patterns of the abbreviation "AG" and related acronyms or significances

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

WARD

-0.79

ndra

-0.77

 stiffness

-0.69

trak

-0.65

 Shutterstock

-0.62

 Spock

-0.61

 agitation

-0.59

asar

-0.59

icycle

-0.58

 imbalance

-0.57

POSITIVE LOGITS

CLE

0.81

lege

0.79

å£«

0.78

ulhu

0.76

Bank

0.75

aneous

0.74

Cola

0.73

ffe

0.72

uder

0.70

onut

0.70

Activations Density 0.233%

proper names or locations in a news context

recurring patterns of the abbreviation "AG" and related acronyms or significances

No Comments

No Known Activations