INDEX

Explanations

questions posed to an audience

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 igno

-0.74

 unif

-0.72

 seiz

-0.72

 unil

-0.72

 lele

-0.69

Okt

-0.64

 grati

-0.63

 roth

-0.62

 sorri

-0.61

 discri

-0.60

POSITIVE LOGITS

How

0.80

how

0.79

why

0.79

 Does

0.78

 whether

0.78

 question

0.77

 Question

0.76

How

0.76

Why

0.76

Does

0.75

Activations Density 0.256%

questions posed to an audience

No Comments

No Known Activations