INDEX

Explanations

phrases related to victories or positive outcomes

oai_token-act-pair · gpt-3.5-turbo

phrases related to winning or success

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

 reminders

-0.66

Ur

-0.66

azel

-0.63

 backlog

-0.62

 Uriel

-0.61

imentary

-0.60

 anten

-0.60

enhagen

-0.60

ugu

-0.59

 ItemLevel

-0.58

POSITIVE LOGITS

win

1.24

Win

0.92

 scenario

0.90

brainer

0.88

 scen

0.81

 proposition

0.81

 outcome

0.80

 loser

0.79

 situation

0.79

win

0.79

Activations Density 0.178%

phrases related to victories or positive outcomes

phrases related to winning or success

No Comments

No Known Activations