INDEX

Explanations

words related to apologies and acts of apologizing

oai_token-act-pair · gpt-3.5-turbo

instances of the word "apologize" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

weeney

-0.88

corn

-0.80

jun

-0.76

spot

-0.76

arnaev

-0.76

marked

-0.75

VEN

-0.71

lining

-0.71

vet

-0.68

eding

-0.68

POSITIVE LOGITS

 apologize

1.14

 apologise

1.08

 apologised

1.07

 apologized

1.07

 apology

1.00

 apologizing

0.99

 apologies

0.97

 sorry

0.84

 unres

0.81

 apolog

0.76

Activations Density 0.012%

words related to apologies and acts of apologizing

instances of the word "apologize" and its variations

No Comments

No Known Activations