INDEX

Explanations

numerical sequences

oai_token-act-pair · gpt-3.5-turbo

numerical values and identifiers related to data or statistics

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

istically

-0.68

 waning

-0.65

anooga

-0.65

 semblance

-0.65

 spons

-0.62

 dramatic

-0.62

 possession

-0.60

 recess

-0.60

 knowledgeable

-0.60

 independence

-0.60

POSITIVE LOGITS

1.79

1.77

1.75

1.66

1.65

1.64

1.62

Activations Density 0.076%

numerical sequences

numerical values and identifiers related to data or statistics

No Known Activations