INDEX

Explanations

references to responsibilities or actions of a specific group or entity

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

 Eleven

-0.75

Posts

-0.70

 Amen

-0.67

 Globe

-0.66

Writ

-0.65

paragraph

-0.64

 Innocent

-0.63

Reviewer

-0.61

 Vanity

-0.60

 Panc

-0.60

POSITIVE LOGITS

're

1.47

've

1.31

'll

1.13

'd

1.06

 were

0.94

 intend

0.92

selves

0.89

are

0.89

 succeeded

0.86

 could

0.86

Activations Density 0.217%

references to responsibilities or actions of a specific group or entity

No Comments

No Known Activations