INDEX

Explanations

phrases or sentences indicating something is useful or beneficial to people

oai_token-act-pair · gpt-3.5-turbo

instances of the word "helpful" and related vocabulary emphasizing assistance or support

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Dare

-0.73

Bam

-0.73

Vision

-0.70

Metro

-0.70

é¾

-0.68

origin

-0.68

thur

-0.67

anyon

-0.65

jong

-0.65

onde

-0.65

POSITIVE LOGITS

aid

0.82

 helpful

0.78

tip

0.77

 aide

0.77

 aids

0.75

glers

0.75

Desk

0.74

 introdu

0.69

 advice

0.69

 tips

0.67

Activations Density 0.020%

phrases or sentences indicating something is useful or beneficial to people

instances of the word "helpful" and related vocabulary emphasizing assistance or support

No Comments

No Known Activations