INDEX

Explanations

phrases relating to holding onto or retaining memories and experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

edback

-0.17

anship

-0.15

ewe

-0.15

ãĥ³ãĥķ

-0.15

tmpl

-0.14

 sticking

-0.14

acts

-0.14

ÑĤÑĮ

-0.14

edy

-0.14

icks

-0.14

POSITIVE LOGITS

ruÅ¾

0.17

 Lester

0.15

exp

0.15

pline

0.14

hold

0.14

IST

0.14

apos

0.14

 hold

0.14

.ke

0.14

ÄĽr

0.13

Activations Density 0.015%

phrases relating to holding onto or retaining memories and experiences

No Comments

No Known Activations