INDEX

Explanations

phrases related to fundraising events for cancer research

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

HuggingFaceFW/fineweb

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disarm

-1.00

<bos>

-0.97

 overwhel

-0.94

 underval

-0.91

 banish

-0.91

 denounce

-0.91

 renounced

-0.90

 repress

-0.89

 usurp

-0.86

 abolish

-0.85

POSITIVE LOGITS

 Research

1.73

Research

1.42

 RESEARCH

1.24

Res

1.24

 catég

1.20

 écout

1.15

 perpétu

1.14

 dégust

1.11

 télécharge

1.11

 plong

1.09

Activations Density 0.078%

phrases related to fundraising events for cancer research

No Comments

No Known Activations