INDEX

Explanations

sports-related terms indicating victory or defeat

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

NR

-0.16

tv

-0.15

Nug

-0.14

odge

-0.14

missible

-0.14

NR

-0.14

Ã«n

-0.14

enha

-0.14

ouncer

-0.14

esso

-0.13

POSITIVE LOGITS

ogh

0.16

undaki

0.15

ICES

0.15

eyin

0.15

HOH

0.14

ices

0.14

leon

0.14

quiet

0.13

 Kemp

0.13

frei

0.13

Activations Density 0.046%

sports-related terms indicating victory or defeat

No Comments

No Known Activations