INDEX

Explanations

phrases indicating scrutiny or examination

oai_token-act-pair · gpt-3.5-turbo

instances of the word "see" indicating observational context or examination

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

shaw

-0.68

inqu

-0.60

alach

-0.57

verse

-0.56

Unknown

-0.56

bey

-0.55

onte

-0.55

pour

-0.55

joice

-0.54

Torrent

-0.54

POSITIVE LOGITS

if

1.49

 whether

1.35

how

1.31

 what

1.16

whether

1.09

 whence

1.04

 firsthand

1.04

why

0.99

 whats

0.96

 where

0.94

Activations Density 0.103%

phrases indicating scrutiny or examination

instances of the word "see" indicating observational context or examination

No Comments

No Known Activations