INDEX

Explanations

scientific and research-related terms

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

holders

-0.71

 Yanuk

-0.67

 disorderly

-0.67

veland

-0.64

 chant

-0.64

erous

-0.62

loaded

-0.61

 timers

-0.61

 servings

-0.60

entity

-0.58

POSITIVE LOGITS

sonian

0.92

 resear

0.91

Researchers

0.90

 researching

0.89

 scientist

0.89

 Researchers

0.86

 Scientist

0.85

 labs

0.83

rador

0.83

 researcher

0.83

Activations Density 4.288%

scientific and research-related terms

No Comments

No Known Activations