INDEX

Explanations

phrases related to specific individuals and leadership positions

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "Os" in relation to 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

 grounds

-0.81

ãĥ¼ãĥĨãĤ£

-0.74

sburgh

-0.71

theless

-0.67

++++

-0.66

 Pilgrim

-0.62

 Dickinson

-0.62

ered

-0.60

ILLE

-0.59

ext

-0.59

POSITIVE LOGITS

hiba

1.16

wered

1.09

boro

0.99

iris

0.99

ugi

0.95

atell

0.92

ound

0.92

awa

0.89

omal

0.89

ophy

0.89

Activations Density 0.011%

phrases related to specific individuals and leadership positions

instances of the abbreviation "Os" in relation to various contexts

No Comments

No Known Activations