INDEX

Explanations

phrases related to nuclear weaponry and national security

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

 âĢº

-0.72

ammy

-0.72

accompan

-0.68

èĢħ

-0.63

JA

-0.61

JP

-0.60

rought

-0.60

veland

-0.60

rir

-0.59

amaz

-0.59

POSITIVE LOGITS

 nothing

1.03

 nobody

0.95

none

0.95

nothing

0.91

 untouched

0.87

None

0.87

 none

0.86

nob

0.85

 unpop

0.84

 remaining

0.83

Activations Density 0.124%

phrases related to nuclear weaponry and national security

No Comments

No Known Activations