INDEX

Explanations

locations and geographical entities

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

ãĥ¼ãĥĨãĤ£

-0.66

ËĪ

-0.56

-0.55

Reviewer

-0.54

ãĤ§

-0.52

ãĥĦ

-0.51

norm

-0.50

adapt

-0.49

ãĥį

-0.49

lim

-0.49

POSITIVE LOGITS

 respectively

1.36

etc

1.29

0.97

*.

0.91

+.

0.87

âĢ¦.

0.84

0.82

etc

0.80

 attRot

0.79

.).

0.79

Activations Density 2.395%

locations and geographical entities

No Comments

No Known Activations