INDEX

Explanations

specific years and dates in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

Ø¯Ø§ÙĨÙĦÙĪØ¯

-0.16

sca

-0.16

erglass

-0.14

PCP

-0.14

ãĥ¼ãĤº

-0.14

bedo

-0.14

ipi

-0.14

 putas

-0.14

ucha

-0.14

uegos

-0.14

POSITIVE LOGITS

uly

0.16

0.15

arest

0.15

0.14

bers

0.14

Activations Density 0.085%

specific years and dates in various contexts

No Comments

No Known Activations