INDEX

Explanations

phrases questioning or exploring reasons for certain occurrences

oai_token-act-pair · gpt-3.5-turbo

expressions of curiosity or inquiries into reasons

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

aughed

-0.85

iece

-0.82

lator

-0.80

 pione

-0.80

đ

-0.77

vertisement

-0.76

rawdownload

-0.76

Ă

-0.76

û

-0.76

ø

-0.76

POSITIVE LOGITS

soever

0.98

 they

0.94

 exactly

0.84

 people

0.82

we

0.80

 nobody

0.79

 someone

0.79

 somebody

0.76

 there

0.75

 anyone

0.73

Activations Density 0.049%

phrases questioning or exploring reasons for certain occurrences

expressions of curiosity or inquiries into reasons

No Comments

No Known Activations