INDEX

Explanations

phrases related to the concept of everything or all-encompassing statements

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ahime

-0.84

edia

-0.71

qus

-0.70

vernment

-0.63

interstitial

-0.62

eka

-0.62

erenn

-0.62

rique

-0.60

[|

-0.60

Trend

-0.59

POSITIVE LOGITS

 except

1.29

 crammed

0.84

 alike

0.83

except

0.82

 equally

0.82

 together

0.82

 toget

0.82

together

0.81

 interconnected

0.77

 individually

0.75

Activations Density 1.783%

phrases related to the concept of everything or all-encompassing statements

No Comments

No Known Activations