INDEX

Explanations

references to violent and criminal incidents involving physical harm

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

 cures

-0.75

 dictators

-0.75

TPP

-0.71

 conscience

-0.70

 divest

-0.69

 optim

-0.69

nuclear

-0.67

 sacrificed

-0.67

 releg

-0.67

 sacrific

-0.66

POSITIVE LOGITS

 Route

0.94

 Cascade

0.90

 Pleasant

0.90

 Northeast

0.89

Smy

0.89

 Meadow

0.87

 Lynn

0.86

 Hampton

0.85

 Interstate

0.84

 Beaver

0.83

Activations Density 0.481%

references to violent and criminal incidents involving physical harm

No Comments

No Known Activations