INDEX

Explanations

references to computer screens or displays

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

ETH

-0.84

vernment

-0.75

ortium

-0.72

EE

-0.70

Uz

-0.70

ALLY

-0.68

 whistle

-0.68

APS

-0.67

ENSE

-0.67

ERE

-0.67

POSITIVE LOGITS

tops

1.12

cloth

1.05

antry

0.90

plays

0.89

 screens

0.89

caps

0.88

 scrolling

0.87

grab

0.85

au

0.84

 layouts

0.79

Activations Density 0.766%

references to computer screens or displays

No Comments

No Known Activations