INDEX

Explanations

personal pronouns preceding verbs indicating action or existence

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

 Tours

-0.64

 resorts

-0.64

 Prior

-0.58

 Gaul

-0.58

 Continental

-0.56

Scroll

-0.55

 Lines

-0.54

 Rebell

-0.54

 Verge

-0.53

 Siber

-0.53

POSITIVE LOGITS

 gonna

1.10

ELF

0.96

selves

0.93

been

0.90

pecially

0.88

initely

0.84

worth

0.81

 gotta

0.78

 been

0.78

 okay

0.77

Activations Density 0.748%

personal pronouns preceding verbs indicating action or existence

No Comments

No Known Activations