INDEX

Explanations

elements or references to time, such as years or dates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

.scalablytyped

-0.17

.Atomic

-0.17

ongyang

-0.17

/window

-0.16

.Undef

-0.15

ÙĦÙĤ

-0.15

ìļ±

-0.14

haft

-0.14

lacak

-0.14

czy

-0.14

POSITIVE LOGITS

ens

0.14

olu

0.13

 notes

0.13

æĬ¼

0.13

_decrypt

0.13

 asses

0.13

 heads

0.13

Activations Density 0.005%

elements or references to time, such as years or dates

No Comments

No Known Activations