INDEX

Explanations

references to tanks or tank-related concepts

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

âĸ¬

-0.95

theless

-0.72

 DMCA

-0.69

çīĪ

-0.68

 intrins

-0.67

BOOK

-0.67

————

-0.66

 rendition

-0.63

éĹ

-0.61

Words

-0.61

POSITIVE LOGITS

erness

1.11

ards

1.05

mates

1.00

ard

1.00

agers

0.98

yard

0.98

ers

0.95

tank

0.93

loads

0.90

arded

0.89

Activations Density 0.008%

references to tanks or tank-related concepts

No Comments

No Known Activations