INDEX

Explanations

variety of cryptographic acronyms and legal terms

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate sources or references in a document

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

opausal

-0.70

iated

-0.68

ctors

-0.66

llor

-0.61

ctor

-0.61

peria

-0.60

 fuelled

-0.60

felt

-0.59

 Loans

-0.58

zinski

-0.58

POSITIVE LOGITS

www

1.18

 http

1.12

 least

1.08

 https

1.05

las

0.98

hens

0.92

 Subtle

0.87

 youtube

0.82

onement

0.79

mega

0.78

Activations Density 0.086%

variety of cryptographic acronyms and legal terms

phrases that indicate sources or references in a document

No Comments

No Known Activations