INDEX

Explanations

terms related to a specific topic, possibly related to card games or strategy games

oai_token-act-pair · gpt-3.5-turbo

terminology related to competitive gaming and strategy

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

 Exhibit

-0.76

 Remem

-0.74

 Anthrop

-0.73

 Bhar

-0.73

 Horton

-0.73

Wem

-0.69

 Dawson

-0.68

uphem

-0.67

 Curt

-0.67

 Geneva

-0.67

POSITIVE LOGITS

 aggro

1.34

 matchups

1.18

 decks

1.15

 setups

1.10

 mages

1.10

 compositions

1.09

DPS

1.06

 mage

1.06

 Zerg

1.05

 Protoss

1.05

Activations Density 0.250%

terms related to a specific topic, possibly related to card games or strategy games

terminology related to competitive gaming and strategy

No Comments

No Known Activations