INDEX

Explanations

situations involving seeking assistance or information in various contexts

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

ohyd

-0.62

 Tycoon

-0.60

ixties

-0.60

fox

-0.60

 Boom

-0.60

azeera

-0.59

POP

-0.57

 Parish

-0.57

 Boone

-0.56

 clicked

-0.56

POSITIVE LOGITS

 urgently

0.88

 clarification

0.86

 forgiveness

0.80

 permission

0.76

 HELP

0.74

 pardon

0.73

 answers

0.73

 assurance

0.71

 attention

0.71

 explanation

0.71

Activations Density 0.266%

situations involving seeking assistance or information in various contexts

No Comments

No Known Activations