INDEX

Explanations

mentions of government entities and officials

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Width

-0.69

 Plate

-0.68

 Candle

-0.67

 Heaven

-0.67

CONT

-0.67

Klu

-0.67

 Velocity

-0.66

 Daylight

-0.65

 Ancest

-0.65

 Tribe

-0.65

POSITIVE LOGITS

 agencies

1.08

 shutdown

1.05

 bureaucracy

1.02

wide

0.99

 officials

0.97

 coffers

0.95

 Accountability

0.95

 departments

0.92

 decree

0.90

 subsidized

0.88

Activations Density 1.353%

mentions of government entities and officials

No Comments

No Known Activations