INDEX

Explanations

text related to editorial notes or commentary in a document

oai_token-act-pair · gpt-3.5-turbo

editorial notes or annotations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

sic

-0.71

anwhile

-0.70

etheless

-0.67

pled

-0.58

 reconc

-0.57

 UNCLASSIFIED

-0.57

bom

-0.55

 skelet

-0.55

ichick

-0.53

 remorse

-0.51

POSITIVE LOGITS

 âĢº

0.83

Phys

0.80

 Transcript

0.63

Description

0.61

][

0.61

 Belfast

0.60

 Immun

0.60

 SHARES

0.59

=>

0.58

î

0.57

Activations Density 0.344%

text related to editorial notes or commentary in a document

editorial notes or annotations

No Comments

No Known Activations