INDEX

Explanations

mentions of doubts or uncertainties

oai_token-act-pair · gpt-3.5-turbo

instances of the contraction "don't" or similar negative contractions

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

 Reborn

-0.79

mere

-0.71

 resumes

-0.67

ersed

-0.63

 srfAttach

-0.62

 Kinnikuman

-0.62

oise

-0.62

 Hels

-0.61

 Palest

-0.60

 Verge

-0.59

POSITIVE LOGITS

 think

1.31

 condone

1.31

 know

1.29

 recall

1.16

 foresee

1.16

 believe

1.16

 understand

1.15

 blame

1.13

 wanna

1.13

 recommend

1.12

Activations Density 0.096%

mentions of doubts or uncertainties

instances of the contraction "don't" or similar negative contractions

No Comments

No Known Activations