INDEX

Explanations

the word "withdraw" or its variations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<bos>

-1.03

 reconno

-0.68

Kak

-0.61

 intrigu

-0.60

 endeav

-0.60

 suspic

-0.59

 vainly

-0.59

 Zeke

-0.59

Kak

-0.58

 negoti

-0.58

POSITIVE LOGITS

 withdrawal

1.15

 withdrawn

1.13

 Withdrawal

1.09

 withdraw

1.09

 Withdraw

1.05

withdraw

1.03

 withdrawals

1.01

 withdrew

0.94

 withdrawing

0.94

withdrawal

0.91

Activations Density 0.448%

the word "withdraw" or its variations

No Comments

No Known Activations