INDEX

Explanations

phrases related to action and consequence

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

emouth

-0.77

ENE

-0.75

LINE

-0.73

eger

-0.72

eenth

-0.70

PLIC

-0.65

okÃ©

-0.65

vertising

-0.64

ural

-0.63

IND

-0.63

POSITIVE LOGITS

 alternatively

1.38

 else

1.35

Else

1.26

chard

0.99

acle

0.96

acles

0.96

phans

0.93

 succumb

0.92

 risk

0.91

 otherwise

0.91

Activations Density 0.096%

phrases related to action and consequence

No Comments

No Known Activations