INDEX

Explanations

instances of various durations, specifically half a year or more, expressed in different contexts

oai_token-act-pair · gpt-3.5-turbo

phrases that reference fractions or portions of time and quantity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

abilities

-0.82

values

-0.76

 intrins

-0.73

Trend

-0.68

videos

-0.68

RAW

-0.66

tests

-0.64

services

-0.64

rules

-0.63

groups

-0.63

POSITIVE LOGITS

 dozen

1.62

dozen

1.46

 century

1.38

 million

1.25

 millennium

1.21

 billion

1.14

decade

1.12

century

1.10

million

1.05

 trillion

1.03

Activations Density 0.040%

instances of various durations, specifically half a year or more, expressed in different contexts

phrases that reference fractions or portions of time and quantity

No Comments

No Known Activations