INDEX

Explanations

actions or events that have happened or been completed

oai_token-act-pair · gpt-3.5-turbo

keywords related to notifications, actions taken, and changes in status or conditions

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

éĥ

-0.69

lication

-0.63

irit

-0.61

VICE

-0.59

role

-0.59

uez

-0.58

nesota

-0.57

when

-0.57

ennis

-0.56

"},{"

-0.56

POSITIVE LOGITS

 since

1.43

 lately

1.13

yet

1.11

yet

1.11

since

1.10

 numerous

1.00

 twice

0.89

 hitherto

0.86

 innumerable

0.85

 extensively

0.85

Activations Density 0.625%

actions or events that have happened or been completed

keywords related to notifications, actions taken, and changes in status or conditions

No Comments

No Known Activations