INDEX

Explanations

administrative-related terms and contexts

oai_token-act-pair · gpt-3.5-turbo

references to administrative processes and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ï¸

-0.86

eem

-0.86

âķĲâķĲ

-0.82

 Emin

-0.76

ï¸ı

-0.75

hov

-0.75

 Kinnikuman

-0.74

gdala

-0.73

çīĪ

-0.73

ku

-0.72

POSITIVE LOGITS

 administrative

0.97

 disciplinary

0.88

 procedures

0.88

 instit

0.88

rative

0.87

 discipline

0.83

 administrator

0.82

 procedure

0.81

 Procedure

0.80

 Procedures

0.79

Activations Density 0.008%

administrative-related terms and contexts

references to administrative processes and actions

No Comments

No Known Activations