INDEX

Explanations

instances of the word "bonus."

oai_token-act-pair · gpt-3.5-turbo

references to bonuses and rewards

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.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

ichen

-0.79

URES

-0.78

tics

-0.78

alter

-0.78

apest

-0.77

URRENT

-0.77

turn

-0.77

ãĤ¢ãĥ«

-0.77

atters

-0.75

AB

-0.75

POSITIVE LOGITS

 bonus

1.27

 bonuses

1.13

 depreciation

0.97

 accrued

0.90

 redundancy

0.87

 incentive

0.83

 extras

0.80

 Bonus

0.80

 parachute

0.80

 accol

0.80

Activations Density 0.012%

instances of the word "bonus."

references to bonuses and rewards

No Comments

No Known Activations