INDEX

Explanations

terms related to communication and information sharing

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

steen

-0.90

iless

-0.89

jury

-0.88

lein

-0.86

endon

-0.85

gling

-0.84

ipedia

-0.74

ighed

-0.74

urat

-0.73

engeance

-0.72

POSITIVE LOGITS

 channels

1.01

 networks

0.93

 communications

0.90

communications

0.86

 messages

0.85

 medium

0.85

 communication

0.85

 transmissions

0.85

 protocols

0.83

 channel

0.81

Activations Density 0.035%

terms related to communication and information sharing

No Comments

No Known Activations