INDEX

Explanations

negations and refusals in text

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

New Auto-Interp

Configuration

jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 lele

-0.80

 pommes

-0.74

 Nguy

-0.72

 magazin

-0.70

 rong

-0.69

 vian

-0.67

 pama

-0.66

 pipa

-0.66

 adal

-0.65

 Chinois

-0.65

POSITIVE LOGITS

 shenan

0.70

Fuckin

0.67

Bullshit

0.66

 necessarily

0.66

 philanth

0.66

FTFY

0.63

Cringe

0.62

Ehh

0.62

 unspeak

0.61

desertcart

0.61

Activations Density 0.182%

negations and refusals in text

No Comments

No Known Activations