INDEX

Explanations

names of people or entities being chosen or appointed to leadership positions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "lead" in various contexts

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

atoes

-0.75

pton

-0.74

eon

-0.71

rost

-0.66

rehens

-0.64

ramid

-0.64

aves

-0.63

 tuber

-0.62

iss

-0.62

phis

-0.61

POSITIVE LOGITS

ership

0.93

better

0.87

lead

0.80

gers

0.77

 Leading

0.72

boards

0.71

gest

0.70

 usher

0.70

ahime

0.68

leaders

0.66

Activations Density 0.032%

names of people or entities being chosen or appointed to leadership positions

instances of the word "lead" in various contexts

No Comments

No Known Activations