INDEX

Explanations

common phrases related to providing information or instructions

oai_token-act-pair · gpt-3.5-turbo

references to lists or enumerations of items or concepts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cel

-0.74

iva

-0.72

chard

-0.71

culosis

-0.71

fecture

-0.69

tm

-0.68

wen

-0.67

chery

-0.66

cher

-0.66

WT

-0.66

POSITIVE LOGITS

 basics

1.31

 essentials

1.24

 strengths

1.18

 pros

1.16

 salient

1.11

 advantages

1.09

 reasons

1.08

 latest

1.08

 biggest

1.07

 steps

1.04

Activations Density 0.198%

common phrases related to providing information or instructions

references to lists or enumerations of items or concepts

No Comments

No Known Activations