INDEX

Explanations

terms related to leadership positions or titles

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

enegger

-0.69

ITNESS

-0.68

urses

-0.66

Pwr

-0.66

 Neighbor

-0.66

 Fusion

-0.63

agra

-0.62

gm

-0.62

tan

-0.59

 belly

-0.59

POSITIVE LOGITS

boards

1.12

doms

1.05

board

0.98

esses

0.93

hips

0.86

leader

0.85

strate

0.85

pins

0.85

hip

0.84

stration

0.84

Activations Density 0.043%

terms related to leadership positions or titles

No Comments

No Known Activations