INDEX

Explanations

steps and instructions in a list format

oai_token-act-pair · gpt-3.5-turbo

instructions or steps related to processes

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

 disappro

-0.75

 contradicted

-0.74

 benefited

-0.72

 valued

-0.71

 lasted

-0.68

 longevity

-0.67

 outwe

-0.66

 disputed

-0.66

 diminishing

-0.66

 benef

-0.65

POSITIVE LOGITS

Create

1.14

Step

1.10

Install

1.10

Place

1.07

Firstly

1.07

Open

1.07

 Firstly

1.06

Copy

1.06

Setup

1.05

Fill

1.04

Activations Density 0.216%

steps and instructions in a list format

instructions or steps related to processes

No Comments

No Known Activations