INDEX

Explanations

conversation snippets involving various interactions and comments between individuals

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

 policymakers

-0.81

 Infrastructure

-0.63

 Policies

-0.62

Europe

-0.61

rastructure

-0.61

verning

-0.61

Financial

-0.59

 reliance

-0.58

 governments

-0.58

 principally

-0.58

POSITIVE LOGITS

 fuckin

0.91

 bitch

0.88

 slang

0.86

 pics

0.84

rhy

0.83

 gigg

0.83

;)

0.82

 joke

0.81

????????

0.80

 dude

0.79

Activations Density 1.470%

conversation snippets involving various interactions and comments between individuals

No Comments

No Known Activations