INDEX

Explanations

situations or scenarios with legal and regulatory implications

oai_token-act-pair · gpt-3.5-turbo

sentences that conclude with periods or are declarative in nature

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 inspir

-0.77

 synthes

-0.74

 skelet

-0.73

 dimensional

-0.72

 impressed

-0.70

 intrigue

-0.69

 burst

-0.69

 appra

-0.68

 excited

-0.67

orb

-0.67

POSITIVE LOGITS

 Previously

1.23

 Instead

1.15

 Additionally

1.13

 Otherwise

1.12

 However

1.09

 Currently

1.09

 Critics

1.07

 Restrict

1.07

 Procedures

1.07

 Anyone

1.06

Activations Density 0.526%

situations or scenarios with legal and regulatory implications

sentences that conclude with periods or are declarative in nature

No Comments

No Known Activations