INDEX

Explanations

terms related to caregiving and instances of reprimand

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

ORGE

-0.69

 understatement

-0.67

 Lomb

-0.65

 disson

-0.65

 Cheong

-0.63

Reviewer

-0.62

 Lauder

-0.62

 marrow

-0.61

uyomi

-0.61

 Fasc

-0.61

POSITIVE LOGITS

itting

1.10

itter

1.06

itters

1.05

chool

0.89

erver

0.88

atcher

0.87

ake

0.85

 babys

0.85

abies

0.82

anny

0.82

Activations Density 0.004%

terms related to caregiving and instances of reprimand

No Comments

No Known Activations