INDEX

Explanations

mentions of titles or achievements in sports or competitions

oai_token-act-pair · gpt-3.5-turbo

terms related to championships and titles in a competitive context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

athy

-0.81

alle

-0.75

merce

-0.75

schild

-0.74

 sshd

-0.70

 Debor

-0.68

artment

-0.67

zee

-0.66

 sourcing

-0.66

 Siber

-0.66

POSITIVE LOGITS

 reigning

1.07

 monarch

0.85

Winner

0.77

 winner

0.76

æľ

0.76

monton

0.71

 winners

0.70

 undefeated

0.69

 undis

0.69

¶ħ

0.67

Activations Density 0.015%

mentions of titles or achievements in sports or competitions

terms related to championships and titles in a competitive context

No Comments

No Known Activations