INDEX

Explanations

expressions conveying certainty or confidence in understanding or doing something

oai_token-act-pair · gpt-3.5-turbo

first-person references to knowledge and self-awareness

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

 Alive

-0.66

âĢ¢âĢ¢

-0.62

 Legends

-0.59

 nowhere

-0.58

 Chron

-0.58

mony

-0.58

 awareness

-0.56

 Dram

-0.56

 Fortune

-0.55

 guiActive

-0.55

POSITIVE LOGITS

're

0.83

mean

0.81

 mean

0.75

 fuss

0.74

doing

0.74

 Mean

0.73

/$

0.72

 meant

0.71

 entail

0.70

need

0.69

Activations Density 0.084%

expressions conveying certainty or confidence in understanding or doing something

first-person references to knowledge and self-awareness

No Comments

No Known Activations