INDEX

Explanations

occurrences of the name "Mur."

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

 Hercules

-0.75

 Pixie

-0.71

 Gemini

-0.68

 Fundamental

-0.66

 OECD

-0.65

 Prometheus

-0.64

 Breach

-0.64

 Gauntlet

-0.63

 Union

-0.63

 incidentally

-0.63

POSITIVE LOGITS

mur

1.27

geon

1.10

asaki

1.08

rats

1.07

abit

1.07

nit

1.06

rill

1.05

phys

1.05

anda

1.00

uj

0.99

Activations Density 0.003%

occurrences of the name "Mur."

No Comments

No Known Activations