INDEX

Explanations

references to something being helpful or beneficial

oai_token-act-pair · gpt-3.5-turbo

instances of the word "useful."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Heart

-0.72

 Kard

-0.68

edited

-0.68

eb

-0.67

chal

-0.66

Rush

-0.66

 Origin

-0.65

Hop

-0.64

Vision

-0.64

eds

-0.63

POSITIVE LOGITS

 useful

1.08

 guiActiveUn

0.94

 referen

0.87

 Useful

0.84

 newsp

0.81

 usefulness

0.79

 behavi

0.79

iable

0.78

 glim

0.78

 idiots

0.77

Activations Density 0.007%

references to something being helpful or beneficial

instances of the word "useful."

No Comments

No Known Activations