INDEX

Explanations

conjunctions that connect phrases or clauses together

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

activex

-0.16

anÃ¡

-0.15

 nÃło

-0.14

AndUpdate

-0.14

imos

-0.13

enko

-0.13

nad

-0.13

Gap

-0.13

ugins

-0.13

ymi

-0.13

POSITIVE LOGITS

etc

0.18

etc

0.18

-most

0.17

 even

0.16

rog

0.15

most

0.15

 finally

0.15

 overall

0.15

 hatta

0.15

 importantly

0.14

Activations Density 0.151%

conjunctions that connect phrases or clauses together

No Comments

No Known Activations