INDEX

Explanations

solutions or attempts to solve problems

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

ittal

-0.74

etry

-0.67

prints

-0.67

ared

-0.66

warm

-0.64

 externalToEVAOnly

-0.64

eks

-0.63

Champ

-0.62

soft

-0.62

Zup

-0.62

POSITIVE LOGITS

 deficiencies

1.03

 grievances

0.97

 misconceptions

0.97

 problems

0.94

 boredom

0.91

 pesky

0.90

balance

0.89

 congestion

0.89

 shortcomings

0.89

 misunderstand

0.88

Activations Density 2.344%

solutions or attempts to solve problems

No Comments

No Known Activations