INDEX

Explanations

verbs related to actions and qualities, but with a focus on negative actions and qualities

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

 objective

-0.64

archment

-0.62

 Window

-0.59

 identification

-0.58

 encyclopedia

-0.58

 Nort

-0.58

 adventurous

-0.57

Disclaimer

-0.57

 cooler

-0.57

 belonging

-0.57

POSITIVE LOGITS

ifies

1.44

izes

1.35

ulates

1.25

itates

1.25

wrote

1.25

uates

1.20

poses

1.15

iates

1.14

cedes

1.08

ounced

1.08

Activations Density 1.194%

verbs related to actions and qualities, but with a focus on negative actions and qualities

No Comments

No Known Activations