INDEX

Explanations

statements including a list of names preceded by titles

oai_token-act-pair · gpt-3.5-turbo

the presence of names and credits associated with individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ptions

-0.79

condition

-0.68

=\"

-0.66

 feasible

-0.66

ption

-0.66

âĶĢ

-0.66

activity

-0.65

 lower

-0.65

Standard

-0.65

employment

-0.65

POSITIVE LOGITS

 Alison

1.32

 Kathleen

1.30

 Stephanie

1.28

 Laurie

1.28

 Jennifer

1.28

 Rebecca

1.28

 Danielle

1.28

 Erica

1.27

 Vanessa

1.27

 Deborah

1.27

Activations Density 0.146%

statements including a list of names preceded by titles

the presence of names and credits associated with individuals

No Comments

No Known Activations