INDEX

Explanations

mentions of specific companies followed by positive or negative opinions or actions

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate possession or attributes related to various subjects

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

FL

-0.73

Interested

-0.69

ear

-0.67

oller

-0.67

click

-0.64

upper

-0.64

INK

-0.63

ential

-0.62

NH

-0.61

bt

-0.60

POSITIVE LOGITS

 always

1.30

 undergone

1.13

 existed

1.10

 been

1.09

 ALWAYS

1.04

 perfected

1.02

 mastered

1.01

 traditionally

1.00

 embraced

1.00

 gotten

0.99

Activations Density 0.251%

mentions of specific companies followed by positive or negative opinions or actions

phrases that indicate possession or attributes related to various subjects

No Comments

No Known Activations