INDEX

Explanations

adjectives related to customization or modification

oai_token-act-pair · gpt-3.5-turbo

terms related to adjustability and flexibility in design

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

src

-0.99

ynthesis

-0.88

achu

-0.87

sic

-0.84

ouls

-0.83

ys

-0.82

orr

-0.81

esville

-0.81

keeper

-0.80

insk

-0.80

POSITIVE LOGITS

 reversible

0.98

 wooden

0.92

 removable

0.92

theless

0.92

 adjustable

0.88

 decorative

0.86

 triangular

0.83

 magnetic

0.83

 curved

0.83

 modular

0.83

Activations Density 0.027%

adjectives related to customization or modification

terms related to adjustability and flexibility in design

No Comments

No Known Activations