INDEX

Explanations

words and suffixes that indicate qualities or states of being

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

atron

-0.16

intosh

-0.15

lett

-0.15

èªĮ

-0.14

uropean

-0.14

feld

-0.14

LETTE

-0.14

ContentLoaded

-0.14

éĥİ

-0.14

ãĥ£

-0.13

POSITIVE LOGITS

ones

0.16

Hol

0.15

etros

0.15

ÂĢÂĻ

0.14

sum

0.14

ÐµÑĢÐ¾

0.13

aan

0.13

Hol

0.13

etro

0.13

arged

0.13

Activations Density 0.293%

words and suffixes that indicate qualities or states of being

No Comments

No Known Activations