INDEX

Explanations

dates and terms related to written publications

oai_token-act-pair · gpt-3.5-turbo

references to specific objects or subjects with a focus on the word "its."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĤ©

-0.80

 Cosponsors

-0.71

enegger

-0.71

terior

-0.71

 ãĤµ

-0.70

lvl

-0.68

ãĥ³ãĤ¸

-0.66

å§

-0.65

 Brach

-0.65

 Ezek

-0.65

POSITIVE LOGITS

onyms

0.68

 bonded

0.65

illon

0.63

anchester

0.63

 detector

0.62

 avatar

0.62

guyen

0.62

 detectors

0.60

 mates

0.60

ploy

0.60

Activations Density 0.000%

dates and terms related to written publications

references to specific objects or subjects with a focus on the word "its."

No Comments

No Known Activations