INDEX

Explanations

descriptions of actions and events involving people in various locations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ema

-0.69

phies

-0.66

pora

-0.66

redo

-0.65

ulnerability

-0.63

ogene

-0.62

nea

-0.62

ooting

-0.61

cknowled

-0.60

cia

-0.60

POSITIVE LOGITS

 frantically

0.69

Sov

0.68

Ern

0.67

 eyed

0.65

 furiously

0.64

 exha

0.63

 dangerously

0.63

itored

0.61

RAG

0.61

redients

0.60

Activations Density 0.223%

descriptions of actions and events involving people in various locations

No Comments

No Known Activations