INDEX

Explanations

instances of the letter 'i' across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

akis

-0.20

.pk

-0.15

.connector

-0.15

ãĤĽ

-0.15

 viá»ĩn

-0.15

ÑĢÐ¾Ð¸Ð·

-0.15

DCALL

-0.15

apult

-0.14

ãĥ¼ãĥĩ

-0.14

TickCount

-0.14

POSITIVE LOGITS

 Hyper

0.15

é¡

0.15

0.14

GS

0.14

mad

0.14

Des

0.14

aven

0.14

 hyper

0.14

Des

0.14

ÑĢÐ°Ð»

0.14

Activations Density 0.027%

instances of the letter 'i' across various contexts

No Comments

No Known Activations