INDEX

Explanations

words related to comparisons or contrasts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

hood

-0.70

sat

-0.69

SPONSORED

-0.64

ÂłÂłÂłÂłÂłÂłÂłÂł

-0.63

deen

-0.63

fair

-0.62

sylvania

-0.62

mouth

-0.62

sov

-0.60

she

-0.59

POSITIVE LOGITS

 regards

2.02

 regard

1.90

 respect

1.52

draw

1.47

stood

1.44

standing

1.27

drawn

1.22

holding

1.19

 impunity

1.12

 hindsight

1.03

Activations Density 0.189%

words related to comparisons or contrasts

No Comments

No Known Activations