INDEX

Explanations

words related to philosophical and theoretical concepts

oai_token-act-pair · gpt-3.5-turbo

phrases centered around foundational principles and arguments in reasoning and philosophy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

agues

-0.73

Ð¾Ð

-0.70

enery

-0.70

soon

-0.70

Ð°

-0.70

NPR

-0.69

LAN

-0.68

enium

-0.67

Netflix

-0.67

Crew

-0.67

POSITIVE LOGITS

 notions

1.28

 causation

1.24

 distinguishing

1.13

 asserting

1.11

 assigning

1.10

 invoking

1.09

 recognizing

1.09

 presupp

1.07

 methodological

1.05

 considerations

1.05

Activations Density 0.307%

words related to philosophical and theoretical concepts

phrases centered around foundational principles and arguments in reasoning and philosophy

No Comments

No Known Activations