INDEX

Explanations

numbers and proper nouns related to political events and entities

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

New Auto-Interp

Configuration

jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.6.hook_resid_post_16384_anthropic_fast_lr

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

torch.float32

Hook Point

blocks.6.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

<bos>

-3.56

 intersper

-2.04

 shenan

-1.71

 apprehen

-1.50

 indestru

-1.45

 hentai

-1.45

 pamph

-1.43

 reluct

-1.43

 inconce

-1.42

 unspeak

-1.42

POSITIVE LOGITS

uhr

1.07

 kado

1.06

 hunde

1.00

 benzina

0.99

 kosme

0.95

 nabi

0.95

 zoll

0.93

 combusti

0.89

 kalori

0.89

 minimalis

0.89

Activations Density 4.784%

numbers and proper nouns related to political events and entities

No Comments

No Known Activations