INDEX

Explanations

references to hierarchical structures or positions within a group or organization

oai_token-act-pair · gpt-3.5-turbo

the term "ranks" in various contexts, particularly relating to hierarchical or organizational structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

argo

-0.99

\/\/

-0.69

vous

-0.69

oll

-0.63

 Archangel

-0.63

uras

-0.61

adra

-0.61

ayne

-0.61

absor

-0.61

 Robo

-0.60

POSITIVE LOGITS

 ranks

1.36

 rank

0.94

 ranked

0.83

Rank

0.80

 veter

0.79

ularity

0.77

 rankings

0.77

masters

0.77

chool

0.73

staff

0.73

Activations Density 0.005%

references to hierarchical structures or positions within a group or organization

the term "ranks" in various contexts, particularly relating to hierarchical or organizational structures

No Comments

No Known Activations