INDEX

Explanations

terminology related to impactful and critical situations or events, such as "crisis."

oai_token-act-pair · gpt-3.5-turbo

recurring mentions of the term "crisis."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

alty

-0.90

alties

-0.88

ross

-0.85

rica

-0.84

rib

-0.82

rosse

-0.81

collar

-0.81

essee

-0.80

ateur

-0.79

cling

-0.78

POSITIVE LOGITS

 crisis

1.03

 Crisis

0.94

 engulf

0.90

 outbreak

0.83

 crises

0.83

 aver

0.82

 epidemic

0.81

 erupt

0.80

 plag

0.78

 erupted

0.77

Activations Density 0.014%

terminology related to impactful and critical situations or events, such as "crisis."

recurring mentions of the term "crisis."

No Comments

No Known Activations