INDEX

Explanations

references to politicians and administrations, specifically mentioning events or actions related to them

oai_token-act-pair · gpt-3.5-turbo

references to U.S. presidential administrations and their actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Sense

-0.71

Self

-0.70

bleacher

-0.69

 deduction

-0.66

oret

-0.65

 reader

-0.65

TOR

-0.64

Phys

-0.64

Channel

-0.62

orah

-0.62

POSITIVE LOGITS

 administrations

1.05

era

0.98

 predecessors

0.98

 eras

0.97

Era

0.93

 oversaw

0.91

 predecessor

0.89

omics

0.85

 presidency

0.82

 administration

0.81

Activations Density 0.246%

references to politicians and administrations, specifically mentioning events or actions related to them

references to U.S. presidential administrations and their actions

No Comments

No Known Activations