INDEX

Explanations

words related to defending, uncovering secrets, and engaging in adventures

oai_token-act-pair · gpt-3.5-turbo

elements related to heroism and adventurous quests

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

ascript

-0.87

TPPStreamerBot

-0.83

 GOODMAN

-0.77

Discuss

-0.74

Write

-0.73

)))

-0.73

displayText

-0.73

-0.72

uliffe

-0.71

zik

-0.70

POSITIVE LOGITS

 supernatural

1.07

 Shinra

1.01

 mysterious

1.01

 kidnapped

0.99

 paranormal

0.99

 immortality

0.94

Osc

0.93

 revenge

0.92

 mutants

0.92

 nefarious

0.91

Activations Density 0.450%

words related to defending, uncovering secrets, and engaging in adventures

elements related to heroism and adventurous quests

No Comments

No Known Activations