INDEX

Explanations

words related to government programs or policies

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

<?

-0.75

 Cosponsors

-0.74

WER

-0.71

eret

-0.64

lly

-0.63

among

-0.62

 notwithstanding

-0.62

detail

-0.61

 Constantin

-0.60

ãĥ³ãĤ¸

-0.59

POSITIVE LOGITS

 arrive

0.97

 expires

0.94

 arrives

0.94

 arrived

0.88

 reaches

0.83

uberty

0.81

 expire

0.80

 enters

0.79

 began

0.79

 expiration

0.78

Activations Density 5.768%

words related to government programs or policies

No Comments

No Known Activations