INDEX

Explanations

references to various types of resources, documents, and structured data

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

 Relief

-0.68

 Apocalypse

-0.66

 Alone

-0.65

 charism

-0.64

-0.62

 Fury

-0.62

-0.60

 Territories

-0.59

cknow

-0.59

POSITIVE LOGITS

 imaginable

1.08

mith

0.96

 vying

0.93

 besides

0.91

paces

0.91

 including

0.87

 ranging

0.85

 apiece

0.85

hips

0.83

 converge

0.79

Activations Density 0.206%

references to various types of resources, documents, and structured data

No Comments

No Known Activations