INDEX

Explanations

terms related to exchanges or trades between parties

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

ahime

-0.86

vironment

-0.80

meta

-0.78

inventoryQuantity

-0.77

 Epidem

-0.76

spect

-0.76

inem

-0.75

ecycle

-0.73

alach

-0.72

ussen

-0.71

POSITIVE LOGITS

 concessions

1.19

 bribes

1.14

 agreeing

1.06

 assurances

1.02

 forfeit

1.02

 pledges

0.99

 ransom

0.96

 guaranteeing

0.95

 payment

0.94

 guarantees

0.93

Activations Density 0.336%

terms related to exchanges or trades between parties

No Comments

No Known Activations