INDEX

Explanations

terms related to legal or procedural language

oai_token-act-pair · gpt-3.5-turbo

terms and phrases related to processes, methods, and production activities

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

 Tough

-0.74

 dads

-0.71

nov

-0.68

abase

-0.67

Rid

-0.66

Bam

-0.65

 srfAttach

-0.65

 Depression

-0.64

 Warren

-0.64

Sand

-0.63

POSITIVE LOGITS

 knowingly

0.99

 alteration

0.91

 lawfully

0.91

 persons

0.91

 materially

0.91

 undertaking

0.91

 unlawfully

0.89

 dispose

0.89

 disposition

0.88

soever

0.87

Activations Density 0.157%

terms related to legal or procedural language

terms and phrases related to processes, methods, and production activities

No Comments

No Known Activations