INDEX

Explanations

words related to negative actions or descriptions

oai_token-act-pair · gpt-3.5-turbo

the letter 'd'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

éĹĺ

-0.89

 Remem

-0.77

-0.76

terday

-0.70

EStream

-0.70

å§«

-0.70

Nun

-0.68

 constitu

-0.68

ãĥ´ãĤ¡

-0.67

OAD

-0.67

POSITIVE LOGITS

etermin

1.14

idd

1.10

ashes

1.08

ams

1.07

iving

1.06

ivers

1.06

ashing

1.05

ividual

1.05

abb

1.04

ivid

1.03

Activations Density 0.027%

words related to negative actions or descriptions

the letter 'd'

No Comments

No Known Activations