INDEX

Explanations

phrases related to proving oneself or something to others, often with a sense of competition or demonstration

oai_token-act-pair · gpt-3.5-turbo

phrases indicating proof of worth or capability

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

roman

-0.74

cheon

-0.71

anon

-0.69

inately

-0.69

enhagen

-0.64

anamo

-0.64

 Lynn

-0.63

erest

-0.62

catentry

-0.62

ranging

-0.61

POSITIVE LOGITS

 usefulness

0.72

 equival

0.70

 faithful

0.69

 relevance

0.69

 conclusive

0.68

 viability

0.68

 chops

0.66

cellence

0.64

 marqu

0.62

 validity

0.62

Activations Density 0.105%

phrases related to proving oneself or something to others, often with a sense of competition or demonstration

phrases indicating proof of worth or capability

No Comments

No Known Activations