INDEX

Explanations

phrases related to generating and discussing ideas

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

administ

-0.71

Naz

-0.70

[+

-0.69

Veter

-0.67

Thompson

-0.65

ords

-0.65

Aff

-0.64

ORD

-0.63

ORPG

-0.63

aths

-0.62

POSITIVE LOGITS

 Ideas

0.94

 ideas

0.86

ually

0.85

 provoking

0.77

 sugg

0.75

 hatched

0.72

 idea

0.72

 inspir

0.71

 embodied

0.70

reprene

0.70

Activations Density 0.040%

phrases related to generating and discussing ideas

No Comments

No Known Activations