INDEX

Explanations

quantitative descriptions of characteristics or events

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agate

-0.77

oran

-0.73

alis

-0.73

ourses

-0.67

 City

-0.67

 Dynamics

-0.66

oris

-0.66

City

-0.65

 messenger

-0.65

endi

-0.63

POSITIVE LOGITS

 identical

0.90

 unanimous

0.88

 impossible

0.81

 nonexistent

0.79

thood

0.78

 limitless

0.78

 unanimously

0.77

zero

0.77

 certainly

0.76

 indistinguishable

0.75

Activations Density 0.904%

quantitative descriptions of characteristics or events

No Comments

No Known Activations