INDEX

Explanations

instances of the words "and", "go", and "out", indicating a focus on action and conjunction in sentences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 6-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L6R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.6.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

-placeholder

-0.15

aises

-0.15

uing

-0.14

cles

-0.14

cular

-0.14

ãĤģãģ¦

-0.14

cling

-0.14

 Bros

-0.13

posable

-0.13

 backwards

-0.13

POSITIVE LOGITS

winter

0.14

 nich

0.14

¸

0.14

à¸ģà¸£à¸°

0.14

sehen

0.14

appy

0.14

adata

0.13

icap

0.13

 keyed

0.13

laÄį

0.13

Activations Density 0.087%

instances of the words "and", "go", and "out", indicating a focus on action and conjunction in sentences

No Comments

No Known Activations