INDEX

Explanations

phrases indicating significant impacts or effects, especially those that are negative or chilling in nature

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

hooting

-0.74

owship

-0.73

 Lands

-0.64

Reviewer

-0.63

bis

-0.62

 aware

-0.61

 Fighting

-0.60

bys

-0.60

Train

-0.60

estern

-0.59

POSITIVE LOGITS

 implications

1.13

 effect

1.12

 impact

1.10

 reperc

1.07

 ramifications

1.07

 significance

1.06

 repercussions

1.04

 consequences

0.98

 resonance

0.96

 origins

0.96

Activations Density 0.149%

phrases indicating significant impacts or effects, especially those that are negative or chilling in nature

No Comments

No Known Activations