INDEX

Explanations

abstract concepts and ideas related to thoughts, assumptions, and narratives

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

Interstitial

-0.76

poon

-0.76

late

-0.68

vez

-0.67

cale

-0.63

emn

-0.63

NetMessage

-0.63

pine

-0.61

âĹ¼

-0.61

TPPStreamerBot

-0.61

POSITIVE LOGITS

 arises

0.87

 begs

0.81

 arose

0.77

 revolves

0.77

 persists

0.73

 seems

0.72

 assumption

0.72

is

0.70

 remains

0.70

 boils

0.69

Activations Density 0.102%

abstract concepts and ideas related to thoughts, assumptions, and narratives

No Comments

No Known Activations