INDEX

Explanations

instances of the word "can" followed by a statement of possibility or permission

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

livious

-0.81

head

-0.69

 Roller

-0.68

roth

-0.64

 Arri

-0.59

ruary

-0.59

advertisement

-0.58

OV

-0.56

rollment

-0.56

 Revis

-0.56

POSITIVE LOGITS

 practicable

1.14

 reasonably

1.05

necessary

1.01

 feasible

1.00

 desired

1.00

 feas

0.98

 convenient

0.97

needed

0.94

 necessary

0.93

 possible

0.91

Activations Density 1.337%

instances of the word "can" followed by a statement of possibility or permission

No Comments

No Known Activations