INDEX

Explanations

phrases related to communication, correspondence, and advocacy

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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.72

clips

-0.72

ranean

-0.71

ocaust

-0.68

arthy

-0.68

ãĤ¨ãĥ«

-0.67

ãĤ¤ãĥĪ

-0.67

avascript

-0.67

ãĥĻ

-0.66

VERTISEMENT

-0.65

POSITIVE LOGITS

 manufacturers

1.05

 landlords

1.05

 retailers

1.03

 publishers

1.03

 legislators

1.01

 suppliers

1.00

 representatives

1.00

 employers

0.99

 superiors

0.98

 landowners

0.97

Activations Density 1.791%

phrases related to communication, correspondence, and advocacy

No Comments

No Known Activations