INDEX

Explanations

references to historical figures and events

oai_token-act-pair · gpt-3.5-turbo

historical references related to monarchs and royal lineage

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 opioids

-0.78

 opioid

-0.76

Sensor

-0.74

MAT

-0.74

INESS

-0.74

stakes

-0.72

sticks

-0.72

lihood

-0.72

NESS

-0.71

 feedback

-0.71

POSITIVE LOGITS

 Augustus

1.15

XVI

1.09

Tud

1.06

 Napoleon

1.02

 VIII

1.00

XIV

1.00

 Ferdinand

0.96

 Emperor

0.95

 grandson

0.93

 Seym

0.90

Activations Density 0.319%

references to historical figures and events

historical references related to monarchs and royal lineage

No Comments

No Known Activations