INDEX

Explanations

phrases related to cause and effect, especially highlighting direct relationships

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

anners

-0.76

 wisely

-0.72

 cautiously

-0.69

gerald

-0.68

Sleep

-0.67

 mell

-0.66

 safely

-0.65

 Frie

-0.65

glers

-0.63

gd

-0.63

POSITIVE LOGITS

 contradicted

0.84

 contradicts

0.76

 opposite

0.75

oire

0.75

aneously

0.74

ebted

0.73

 contradict

0.73

orship

0.73

aneous

0.73

 sunlight

0.68

Activations Density 1.562%

phrases related to cause and effect, especially highlighting direct relationships

No Comments

No Known Activations