INDEX

Explanations

specific numerical and code-like identifiers in text

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

iol

-0.16

eon

-0.14

bdd

-0.14

Isl

-0.14

erable

-0.14

Ø¨ÙĪØ±

-0.14

/workspace

-0.14

à¤¿à¤ļ

-0.14

iÄħ

-0.13

hower

-0.13

POSITIVE LOGITS

çļĦæĺ¯

0.16

Tru

0.14

Ð»Ð¸ÑĤ

0.14

 Kirk

0.14

æ°´å¹³

0.14

Rx

0.13

fÃ¼

0.13

Ð¾Ð±Ð°

0.13

claimer

0.13

Dispatch

0.13

Activations Density 0.057%

specific numerical and code-like identifiers in text

No Comments

No Known Activations