INDEX

Explanations

phrases related to reasonableness or reason in a context of evaluation or judgment

oai_token-act-pair · gpt-3.5-turbo

the term "reasonable" in various contexts and its different usages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

sten

-0.69

nen

-0.69

fen

-0.68

berman

-0.68

Hop

-0.67

board

-0.67

 Lust

-0.66

Indust

-0.66

chin

-0.65

inis

-0.64

POSITIVE LOGITS

 reasonable

1.01

 accommodation

0.96

asonable

0.94

 accommodations

0.89

 sane

0.89

 expectation

0.88

 compr

0.86

ãĥ¼ãĥĨ

0.85

erate

0.82

 enough

0.82

Activations Density 0.011%

phrases related to reasonableness or reason in a context of evaluation or judgment

the term "reasonable" in various contexts and its different usages

No Comments

No Known Activations