INDEX

Explanations

proper nouns or named entities within personal narratives

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

abal

-0.70

 conclud

-0.67

 consumed

-0.63

imately

-0.63

 Liver

-0.62

edia

-0.61

icip

-0.60

iland

-0.59

 reaches

-0.57

imedia

-0.56

POSITIVE LOGITS

 bluff

1.29

ãĥ¼ãĥ³

0.83

 hotline

0.77

EStream

0.72

Tes

0.71

 Duty

0.69

 shots

0.68

 bullshit

0.68

icide

0.65

 liar

0.65

Activations Density 1.032%

proper nouns or named entities within personal narratives

No Comments

No Known Activations