INDEX

Explanations

terms related to the software program "Microsoft Word"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Word" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

asio

-0.80

olson

-0.78

Interstitial

-0.77

ocal

-0.77

 unnecess

-0.74

avorite

-0.74

alus

-0.72

iery

-0.71

etheus

-0.71

reens

-0.70

POSITIVE LOGITS

sworth

1.13

 Word

1.08

press

1.05

Word

0.88

Press

0.84

Perfect

0.82

 Trafford

0.81

 Bunker

0.75

word

0.75

ifier

0.74

Activations Density 0.007%

terms related to the software program "Microsoft Word"

occurrences of the word "Word" in various contexts

No Comments

No Known Activations