INDEX

Explanations

phrases related to actions involving going somewhere and doing something, often in relation to others

oai_token-act-pair · gpt-3.5-turbo

coordinating conjunctions that indicate conjunction or connection between ideas

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

etheless

-0.85

immune

-0.70

erred

-0.66

ESCO

-0.62

derived

-0.61

drawn

-0.61

ortunately

-0.61

dinand

-0.60

driving

-0.58

ANC

-0.58

POSITIVE LOGITS

 forth

1.04

say

0.88

buy

0.82

 partake

0.82

 grab

0.81

 compete

0.79

 spoil

0.79

try

0.78

 yell

0.78

 whine

0.78

Activations Density 0.130%

phrases related to actions involving going somewhere and doing something, often in relation to others

coordinating conjunctions that indicate conjunction or connection between ideas

No Comments

No Known Activations