INDEX

Explanations

terms related to one-on-one interactions or activities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating interactive one-on-one activities

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

ãĥ¼ãĥĨãĤ£

-0.70

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

-0.70

ynt

-0.62

 Gillespie

-0.59

 crow

-0.56

Sy

-0.55

 McDonnell

-0.55

WER

-0.54

 revised

-0.53

 Timeline

-0.53

POSITIVE LOGITS

ones

0.97

one

0.90

onson

0.85

ouch

0.78

owitz

0.76

two

0.74

ONES

0.73

onet

0.73

ONE

0.73

onis

0.72

Activations Density 0.070%

terms related to one-on-one interactions or activities

phrases indicating interactive one-on-one activities

No Comments

No Known Activations