INDEX

Explanations

references to tax-related terms or concepts

oai_token-act-pair · gpt-3.5-turbo

references to taxes

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

lihood

-0.82

story

-0.73

theless

-0.70

 Liberation

-0.70

 Sadd

-0.68

 Encounter

-0.68

 Condition

-0.67

Voy

-0.64

ciating

-0.64

Found

-0.64

POSITIVE LOGITS

 levied

1.11

 taxes

1.10

heet

1.06

payer

1.04

 payable

1.01

 Taxes

1.00

 owed

0.92

 collectors

0.88

tax

0.86

 taxed

0.86

Activations Density 0.018%

references to tax-related terms or concepts

references to taxes

No Comments

No Known Activations