INDEX

Explanations

phrases that indicate intentions or actions involving obligation or recommendation

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

VERTISEMENT

-0.76

luck

-0.66

washer

-0.64

 srfAttach

-0.64

lore

-0.63

tops

-0.60

requires

-0.60

 Survivors

-0.59

gradation

-0.57

grounds

-0.57

POSITIVE LOGITS

 dedicate

1.01

 pursue

0.98

 undertake

0.97

 choose

0.96

 embrace

0.93

 embark

0.92

 publish

0.92

 participate

0.92

 invest

0.92

 expose

0.91

Activations Density 0.105%

phrases that indicate intentions or actions involving obligation or recommendation

No Comments

No Known Activations