INDEX

Explanations

keywords related to natural disasters, scientific research, and political and economic issues

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 predec

-0.59

 assignment

-0.56

 schooling

-0.56

 freight

-0.55

 flakes

-0.54

 intermediate

-0.54

 prototyp

-0.54

 frontline

-0.53

 allocation

-0.53

 worm

-0.53

POSITIVE LOGITS

 Includes

0.88

 Secondly

0.88

 Therefore

0.85

 Besides

0.84

 Whilst

0.83

 Especially

0.83

 Âł

0.82

 However

0.82

↵↵

0.82

 Also

0.81

Activations Density 5.565%

keywords related to natural disasters, scientific research, and political and economic issues

No Comments

No Known Activations