INDEX

Explanations

references to significant actions or consequences

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

iman

-0.17

pone

-0.16

ulet

-0.15

inte

-0.15

Tal

-0.15

adal

-0.15

ãĥŀãĥ³

-0.14

ynn

-0.14

mey

-0.14

LOB

-0.14

POSITIVE LOGITS

pcf

0.17

lashes

0.16

çĽĽ

0.15

orida

0.14

isis

0.14

 autoload

0.14

unca

0.14

digits

0.14

prov

0.14

okud

0.14

Activations Density 0.009%

references to significant actions or consequences

No Comments

No Known Activations