INDEX

Explanations

phrases related to legal proceedings and arguments

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, particularly commas, in the text

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

hov

-0.71

natureconservancy

-0.68

ileged

-0.66

inar

-0.66

deen

-0.64

Ã§a

-0.64

ãĥ¯ãĥ³

-0.62

imity

-0.61

 guiActiveUnfocused

-0.61

oir

-0.60

POSITIVE LOGITS

 albeit

1.13

but

0.96

 citing

0.95

 concluding

0.95

 noting

0.94

 unsuccessfully

0.90

 although

0.90

 saying

0.86

 prompting

0.85

 stating

0.85

Activations Density 0.310%

phrases related to legal proceedings and arguments

punctuation marks, particularly commas, in the text

No Comments

No Known Activations