INDEX

Explanations

phrases related to criminal activities and investigations

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

animate

-0.59

 underwear

-0.59

Virgin

-0.59

 slaughtered

-0.57

 starvation

-0.56

 bathing

-0.56

 Tycoon

-0.54

 slaughter

-0.54

ru

-0.54

 raping

-0.53

POSITIVE LOGITS

 methodological

0.64

 policymakers

0.60

 broadly

0.60

 academia

0.59

 outlining

0.58

 FOIA

0.58

 summarize

0.58

 publication

0.58

 disingen

0.58

 nonpartisan

0.58

Activations Density 1.399%

phrases related to criminal activities and investigations

No Comments

No Known Activations