INDEX

Explanations

numerical indicators of game periods and scoring sequences

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

Negative Logits

plusplus

-0.16

asti

-0.16

ìķķ

-0.15

andes

-0.15

ernes

-0.14

.travel

-0.14

 Seasons

-0.14

azar

-0.14

chia

-0.14

Dit

-0.14

POSITIVE LOGITS

 frame

0.31

 period

0.30

frame

0.25

_frame

0.24

period

0.24

 stanza

0.24

 inning

0.24

-frame

0.24

.frame

0.23

 Frame

0.23

Activations Density 0.027%

numerical indicators of game periods and scoring sequences

No Known Activations