INDEX

Explanations

word sequences related to technical processes or instructions

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

 Darío

-0.64

 Mónica

-0.57

 Belén

-0.56

 Portugu

-0.55

 Áng

-0.53

 Agustín

-0.51

ngOn

-0.50

 certamen

-0.50

 Julián

-0.50

 Víctor

-0.49

POSITIVE LOGITS

 chrysler

0.78

 skoda

0.73

jbl

0.72

fuj

0.71

 volkswagen

0.70

 shimano

0.68

 lexus

0.67

 mitsubishi

0.67

 vios

0.67

 hilux

0.66

Activations Density 0.856%

word sequences related to technical processes or instructions

No Comments

No Known Activations