INDEX

Explanations

mentions of hotels or accommodations

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

İĭ

-0.79

nir

-0.78

rations

-0.74

nan

-0.73

-0.71

xit

-0.69

 Anarchy

-0.67

nces

-0.67

ername

-0.65

 instr

-0.64

POSITIVE LOGITS

 rooms

1.10

 occupancy

1.03

 room

1.01

 guests

0.99

 hotel

0.98

 Hotel

0.95

 accommodations

0.93

 lobb

0.89

 lobby

0.87

 hotels

0.86

Activations Density 0.044%

mentions of hotels or accommodations

No Comments

No Known Activations