INDEX

Explanations

terms related to correspondence and communication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

istic

-0.15

ustos

-0.15

FromString

-0.15

uen

-0.15

lein

-0.15

erman

-0.14

ampa

-0.14

til

-0.14

Ð»ÐµÐ³

-0.14

er

-0.14

POSITIVE LOGITS

ingly

0.35

ence

0.21

ences

0.21

ents

0.18

ed

0.18

ENCE

0.18

iente

0.17

ientes

0.15

 roughly

0.15

rough

0.15

Activations Density 0.015%

terms related to correspondence and communication

No Comments

No Known Activations