INDEX

Explanations

punctuation marks following clauses

oai_token-act-pair · gpt-3.5-turbo

comas and phrases indicating continuations or elaborations on previously stated ideas

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

ode

-0.61

iasm

-0.56

eers

-0.56

CHA

-0.55

Score

-0.54

Reward

-0.53

igers

-0.53

sth

-0.52

ODE

-0.52

è¦ļéĨĴ

-0.51

POSITIVE LOGITS

 unlike

1.20

 contrary

1.09

 despite

1.09

 although

1.06

despite

1.05

 irrespective

0.98

 regardless

0.96

 barring

0.96

 whereas

0.95

 insofar

0.94

Activations Density 0.095%

punctuation marks following clauses

comas and phrases indicating continuations or elaborations on previously stated ideas

No Comments

No Known Activations