INDEX

Explanations

references to various types of weapons

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

gres

-0.75

 Strawberry

-0.75

dit

-0.74

cess

-0.69

 Debor

-0.69

lus

-0.68

 Manz

-0.68

Cow

-0.66

 coli

-0.65

 Hort

-0.65

POSITIVE LOGITS

powder

1.10

 wielded

0.97

 salute

0.87

 weapon

0.83

 arsenal

0.83

 manufacturer

0.82

 weapons

0.82

usable

0.79

istics

0.77

 unlocks

0.76

Activations Density 0.018%

references to various types of weapons

No Comments

No Known Activations