INDEX

Explanations

adjectives that express various ways by which things can be done or achieved

oai_token-act-pair · gpt-3.5-turbo

phrases that denote methods or approaches to achieve specific outcomes or goals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

halla

-0.72

INTON

-0.63

dfx

-0.63

endas

-0.63

aan

-0.62

erald

-0.62

osponsors

-0.62

herent

-0.61

igi

-0.61

 Niet

-0.60

POSITIVE LOGITS

 whereby

0.93

fare

0.87

chery

0.76

terday

0.74

 employed

0.73

icles

0.69

finder

0.68

 necessary

0.66

of

0.66

piece

0.66

Activations Density 0.021%

adjectives that express various ways by which things can be done or achieved

phrases that denote methods or approaches to achieve specific outcomes or goals

No Comments

No Known Activations