INDEX

Explanations

statements or phrases asserting the presence or significance of a subject or event

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

ngth

-0.75

styles

-0.70

cheat

-0.68

 floats

-0.68

 guides

-0.65

onomous

-0.64

objects

-0.63

)</

-0.63

fights

-0.63

 slightest

-0.62

POSITIVE LOGITS

 unsur

0.81

 remarkable

0.74

 contrasted

0.74

 hardly

0.73

 reminiscent

0.73

 sympt

0.71

quickShipAvailable

0.69

 unclear

0.68

 apparently

0.67

 ironic

0.66

Activations Density 0.101%

statements or phrases asserting the presence or significance of a subject or event

No Comments

No Known Activations