INDEX

Explanations

actions or tasks that someone can do

oai_token-act-pair · gpt-3.5-turbo

phrases expressing capabilities or actions

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

ONSORED

-0.69

theless

-0.66

lights

-0.66

 sentenced

-0.64

laus

-0.62

 Canal

-0.62

Hung

-0.60

zilla

-0.59

Tai

-0.59

hattan

-0.59

POSITIVE LOGITS

omething

0.91

omsday

0.89

pez

0.81

 anything

0.81

hing

0.77

ggy

0.77

xa

0.76

xx

0.74

ozy

0.74

xy

0.74

Activations Density 0.039%

actions or tasks that someone can do

phrases expressing capabilities or actions

No Comments

No Known Activations