INDEX

Explanations

phrases related to expressing personal opinions or positions

oai_token-act-pair · gpt-3.5-turbo

instances of personal opinion or commentary

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Pric

-0.74

 matching

-0.68

cellence

-0.62

venge

-0.60

 rejuven

-0.60

 honoured

-0.59

 Selection

-0.59

 surpass

-0.59

 disinfect

-0.58

 surpassed

-0.58

POSITIVE LOGITS

say

1.11

 saying

1.04

 said

0.98

mean

0.96

said

0.90

SAY

0.89

say

0.88

Say

0.87

 Said

0.87

 Saying

0.85

Activations Density 0.152%

phrases related to expressing personal opinions or positions

instances of personal opinion or commentary

No Comments

No Known Activations