INDEX

Explanations

terms related to artworks, specifically focused on dimensions or cataloguing styles

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

ORED

-0.81

wagen

-0.74

ãĥ¼ãĥĨãĤ£

-0.67

ously

-0.67

ãģ®éŃĶ

-0.66

arding

-0.65

IFIED

-0.65

DERR

-0.65

 guiActiveUnfocused

-0.64

 administrations

-0.64

POSITIVE LOGITS

mbol

1.48

ntax

1.38

nerg

1.20

thia

1.12

nergy

1.12

racuse

1.09

ptoms

1.09

posium

1.07

philis

1.04

rians

1.04

Activations Density 0.008%

terms related to artworks, specifically focused on dimensions or cataloguing styles

No Comments

No Known Activations