INDEX

Explanations

references to ordinal placements or sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 15-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L15R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.15.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

etimes

-0.16

'].$

-0.15

ÑĪÐ¾Ð²

-0.15

Ð°Ð½ÑģÐ¾Ð²

-0.14

oret

-0.14

adden

-0.14

imler

-0.14

idis

-0.14

arel

-0.14

ìłĢ

-0.14

POSITIVE LOGITS

 part

0.20

 portion

0.18

portion

0.18

 half

0.17

ched

0.16

¿ł

0.16

 portions

0.16

posit

0.15

 Curtis

0.15

ilk

0.15

Activations Density 0.085%

references to ordinal placements or sequences

No Comments

No Known Activations