INDEX

Explanations

references to specific parts of a phone or technology-related articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

perse

-0.84

aneers

-0.82

aughs

-0.80

obia

-0.76

LOD

-0.75

ellig

-0.75

arcity

-0.71

APH

-0.71

ventus

-0.71

FF

-0.70

POSITIVE LOGITS

 underside

1.23

 periphery

1.16

 occasion

1.10

 forehead

1.08

 doorstep

1.07

 shelf

1.06

 bottom

1.06

 horizon

1.00

 wearer

0.98

 floor

0.98

Activations Density 0.147%

references to specific parts of a phone or technology-related articles

No Comments

No Known Activations