INDEX

Explanations

dialogues in a legal or courtroom context

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

etheless

-0.77

 sprung

-0.69

 pellets

-0.67

 care

-0.65

 outwe

-0.65

``

-0.65

 drinkers

-0.65

aband

-0.65

 fest

-0.64

eatures

-0.63

POSITIVE LOGITS

 Unable

1.21

 Same

1.12

 Finally

1.06

 Lastly

1.05

 Attempt

1.05

 Beginning

1.04

 Interesting

1.03

 Beware

1.02

 Whilst

1.02

 Thousands

1.02

Activations Density 1.715%

dialogues in a legal or courtroom context

No Comments

No Known Activations