INDEX

Explanations

references to surprise and disbelief in various contexts

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

 mediation

-0.78

aceutical

-0.68

 Procedures

-0.61

 represents

-0.61

emouth

-0.61

 latitude

-0.61

 maintenance

-0.60

atum

-0.59

 autonomy

-0.59

tesy

-0.59

POSITIVE LOGITS

 laughing

1.15

 shudder

1.14

 envy

1.13

 chuckle

1.10

 cringe

1.09

 gasp

1.07

 amaz

1.02

 incred

1.01

 horrified

1.00

 blush

1.00

Activations Density 0.381%

references to surprise and disbelief in various contexts

No Comments

No Known Activations