INDEX

Explanations

phrases related to user interaction or control within a game or digital environment

oai_token-act-pair · gpt-3.5-turbo

references to roles or identities of individuals in a gaming or interactive context

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

hots

-0.81

Ø§ÙĦ

-0.77

arus

-0.75

poons

-0.72

bows

-0.71

hift

-0.68

hops

-0.68

uca

-0.65

hooting

-0.64

poon

-0.64

POSITIVE LOGITS

's

1.04

 himself

0.99

ÃŃs

0.91

agonist

0.88

osphere

0.87

 Himself

0.85

liest

0.82

abase

0.82

 herself

0.78

ele

0.78

Activations Density 0.176%

phrases related to user interaction or control within a game or digital environment

references to roles or identities of individuals in a gaming or interactive context

No Comments

No Known Activations