INDEX

Explanations

negative or problematic aspects or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

artney

-0.68

essage

-0.64

 Niet

-0.61

Keefe

-0.61

Ó

-0.60

ILA

-0.59

 Barrett

-0.59

 Reilly

-0.58

ppard

-0.58

swick

-0.58

POSITIVE LOGITS

][

0.65

ness

0.62

 enough

0.61

nesses

0.59

 subreddits

0.56

 quotes

0.52

)]

0.52

 disclaimer

0.52

lycer

0.52

Activations Density 0.274%

negative or problematic aspects or situations

No Comments

No Known Activations