INDEX

Explanations

statements or claims related to recent events or news

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 groove

-0.78

 friendships

-0.72

 adolescence

-0.70

ahime

-0.70

 satell

-0.70

 affili

-0.69

 resil

-0.69

 ensemble

-0.69

 nurture

-0.69

 nurt

-0.69

POSITIVE LOGITS

 Therefore

1.43

 Presumably

1.41

 Apparently

1.40

 Consequently

1.33

 Meaning

1.30

 Thus

1.30

Therefore

1.30

 Needless

1.29

 Similarly

1.26

 Furthermore

1.25

Activations Density 2.625%

statements or claims related to recent events or news

No Comments

No Known Activations