INDEX

Explanations

proper nouns and specific targets, especially related to military operations

oai_token-act-pair · gpt-3.5-turbo 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

ansk

-0.70

 Expedition

-0.67

maid

-0.66

IGH

-0.66

ooks

-0.65

©¶æ

-0.63

loo

-0.62

oline

-0.62

ModLoader

-0.61

aut

-0.60

POSITIVE LOGITS

ted

1.26

ting

0.86

 range

0.78

 target

0.76

topic

0.75

ivity

0.74

 targets

0.73

rod

0.72

izen

0.72

ched

0.72

Activations Density 0.028%

proper nouns and specific targets, especially related to military operations

No Comments

No Known Activations