INDEX

Explanations

phrases related to measures aimed at restoring or fixing something that is considered to be damaged or in a negative state

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

gom

-0.76

kers

-0.72

kson

-0.67

gerald

-0.66

ks

-0.65

zag

-0.63

fork

-0.62

Hash

-0.61

QB

-0.61

wer

-0.61

POSITIVE LOGITS

 sanity

0.89

 composure

0.85

vig

0.79

 equilibrium

0.78

 harmony

0.78

 restore

0.76

 balance

0.74

 restoration

0.74

 vitality

0.74

 calm

0.74

Activations Density 0.027%

phrases related to measures aimed at restoring or fixing something that is considered to be damaged or in a negative state

No Comments

No Known Activations