INDEX

Explanations

numbers in a specific format

oai_token-act-pair · gpt-3.5-turbo

numerical values, specifically related to a particular code or identifier format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

purpose

-0.79

mble

-0.78

ellation

-0.73

ende

-0.72

uran

-0.72

nesday

-0.71

alty

-0.71

hold

-0.70

endum

-0.69

sonian

-0.69

POSITIVE LOGITS

 ILCS

1.09

0.92

0.84

0.83

0.82

«ĺ

0.82

ipop

0.82

mph

0.81

0.80

Activations Density 0.023%

numbers in a specific format

numerical values, specifically related to a particular code or identifier format

No Comments

No Known Activations