INDEX

Explanations

specific information related to news articles or reports, including names of individuals, locations, and 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_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

utenberg

-0.80

å¥³

-0.77

 Diss

-0.75

binding

-0.71

 pores

-0.70

ppm

-0.68

wu

-0.68

 Diary

-0.67

parency

-0.67

ipher

-0.66

POSITIVE LOGITS

 vehicles

1.58

 cars

1.56

 vehicle

1.48

 motorcycles

1.46

 passenger

1.45

 automobiles

1.41

 taxis

1.41

 sedan

1.38

SUV

1.36

wagon

1.36

Activations Density 1.822%

specific information related to news articles or reports, including names of individuals, locations, and events

No Comments

No Known Activations